CARTOGRAPHIC ANALYSIS OF POPULATIONS IN CHICAGO
By
Nicholas A. Perdue

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Geography
2012

ABSTRACT
CARTOGRAPHIC ANALYSIS OF POPULATIONS IN CHICAGO
By
Nicholas A. Perdue
The goal of this project is to make the best maps to show how people are concentrated in
urban spaces. The conventional methods of measuring density of populations in cities rely on
calculations of quantities of people living within a surface boundary and fail to account for the
multiple floor residential patterns of the contemporary urban landscape. Conventional population
density measurements do not characterize the crowdedness or spaciousness of the lived
experience of people within cities. To create an accurate representation of people in
contemporary urban spaces, a move beyond the conventional conception of density is needed.
This research aims to find a more appropriate solution to mapping humans in cities by employing
a dasymetric method to represent the distribution of people in a city of vertical residential
structures. The methodology creates an index to classify the amount of floor space for each
person across the extent of the city, a metric called personal space. This measure of personal
space is juxtaposed with the conventional population density measurements to provide a unique
perspective on how population is concentrated across the urban space. Demographic analysis of
the new metric illustrates how conventional approaches misrepresent the urban social landscape.
Chicago, with a large population and a high vertical extent, makes an ideal case study to develop
a methodology that can be applied to all cities. The primary focus of this research is to illustrate
how the conventional measurement of population density fails to capture the phenomena of
urban living in the 21st century and to provide alternative approaches to accurately and
intelligently analyze the contemporary urban space.

ACKNOWLEDGEMENTS

There are many individuals I would like to thank for the support, advice, and direction
during the writing of this thesis. First my advisor, Dr. Kirk Goldsberry, for being patient as I
worked through the various problems and challenges I encountered during the past two years.
Thank you for teaching me how to make maps, for providing a mixture of advice and
encouragement, and for helping stay passionate about the process. I would also like to thank my
committee, Dr. Ashton Shortridge and Dr. Igor Vojnovic for teaching me how to approach my
research questions and providing timely criticism when I needed it.
I would like to thank the Geography Department at Michigan State University for
providing me with the opportunity to work on this research and supporting me in travel to
present my work. The faculty here has always been available and offered great advice and
encouragement. I would also like to thank all my fellow graduate students- I am truly thankful
for being surrounded by a supportive and intelligent group of people the past two years.
Funding for this research was provided by the MSU Department of Geography
Graduate Office Fellowship, the Master Thesis Grant from the Cartography Specialty Group of
the Association of American Geographers, and the North American Cartographic Information
Society.
I would also like to thank all my friends in Denver for support in every aspect of my life
while I have been away. Specifically, my deepest appreciation to Michael Vallee, Megan Kester,
Luke Wacther, Liz Otero, Claire Sheridan, and Jessica Miller. Thank you for everything.
Finally, I would like to thank my parents, my sister Lauren, and all my family for the
support and love through my successes and failures leading to this point. You have been an
inspiration and guiding force throughout my life and I struggle to find the words to express my
gratitude.

iii

TABLE OF CONTENTS

LIST OF TABLES ..................................................................................................................... vi
LIST OF FIGURES ..................................................................................................................vii
Chapter 1
INTRODUCTION ...................................................................................................................... 1
1.1 Introduction ..................................................................................................................... 1
1.2 Research Goals ................................................................................................................ 5
1.3 Research Questions ......................................................................................................... 6
Chapter 2
BACKGROUND ........................................................................................................................ 7
2.1 Introduction ..................................................................................................................... 7
2.2 United States Census ....................................................................................................... 7
2.2.1 Census Background .................................................................................................. 7
2.2.2 Importance of Census ............................................................................................. 10
2.3 Population Density ........................................................................................................ 11
2.3.1 Population Density Overview................................................................................ 11
2.3.2 Alternative Density Measurements ........................................................................ 11
2.3.3 Population Density Limitations ............................................................................. 13
2.3.4 Population Density Models and Applications ........................................................ 14
2.4 Cartographic Representation .......................................................................................... 16
2.4.1 Introduction .......................................................................................................... 16
2.4.2 Choropleth Maps .................................................................................................. 16
2.4.3 Choropleth Maps and Population Density ............................................................. 18
2.4.4 Dot Density Maps ................................................................................................. 19
2.4.5 Dot Density Considerations................................................................................... 20
2.4.6 Dasymetric Maps .................................................................................................. 21
2.4.7 Dasymetric Methods ............................................................................................. 22
2.4.8 Binary Method ...................................................................................................... 23
2.4.9 Areal Interpolation Method ................................................................................... 24
2.4.10 Three Class Method ............................................................................................ 25
2.4.11 Limiting Variable Method ................................................................................... 26
2.4.12 Limitations in Dasymetric Mapping .................................................................... 27
2.4.13 Cartographic Representation Summary ............................................................... 28
2.5 Literature Review Conclusion........................................................................................ 28
Chapter 3
STUDY AREA ......................................................................................................................... 30
3.1 Chicago Introduction ..................................................................................................... 30
3.2 Boundaries in Chicago ................................................................................................... 31

iv

3.3 Growth of Chicago ........................................................................................................ 32
3.4 Community Area Boundary System ............................................................................... 35
3.5 Modern Urban Structure of Chicago .............................................................................. 40
3.6 Population Density in Chicago ....................................................................................... 41
3.7 Chicago Summary ......................................................................................................... 43
Chapter 4
METHODS ............................................................................................................................... 44
4.0 Data............................................................................................................................... 44
4.1 Overview and Goals ...................................................................................................... 45
4.2 Personal Space Model.................................................................................................... 46
4.3 Classification of Personal Space .................................................................................... 49
4.4 Demographic Profiles of Chicago .................................................................................. 51
4.5 Individual Representation Method ................................................................................. 52
4.6 Summary ....................................................................................................................... 53
Chapter 5
RESULTS ................................................................................................................................. 54
5.1 Overview and Goals ...................................................................................................... 54
5.2 Personal Space Model.................................................................................................... 54
5.3 Classification of Personal Space .................................................................................... 57
5.4 Demographic Profiles of Chicago .................................................................................. 67
5.5 Individual Representation Results .................................................................................. 77
5.6 Intensity Results ............................................................................................................ 86
Chapter 6
DISCUSSION ........................................................................................................................... 88
6.1 Overview and Goals ...................................................................................................... 88
6.2 Personal Space Metric ................................................................................................... 88
6.3 Classification of Personal Space .................................................................................... 92
6.4 Demographic Profiles of Chicago .................................................................................. 93
6.5 Individual Representation ........................................................................................... 104
6.6 Contribution to Geography .......................................................................................... 107
Chapter 7
CONCLUSION ....................................................................................................................... 120
7.1 Conclusion .................................................................................................................. 120
REFERENCES ....................................................................................................................... 122

v

LIST OF TABLES

Table 1. Green community areas ............................................................................................... 61
Table 2. Pink community areas.................................................................................................. 62
Table 3. Orange community areas ............................................................................................. 63
Table 4. Purple community areas............................................................................................... 64
Table 5. Demographic Profile of all spaces in Chicago ............................................................. 76

vi

LIST OF FIGURES

Figure 1. Map of Chicago community area regions.................................................................... 38
Figure 2. Choropleth map of population density in Chicago ...................................................... 42
Figure 3. Personal Space Flowchart ........................................................................................... 47
Figure 4. Personal space equation .............................................................................................. 49
Figure 5. Bivariate legend ......................................................................................................... 50
Figure 6. Map of buildings over seven stories in Chicago, 2008 ................................................ 55
Figure 7. Bivariate legend ......................................................................................................... 58
Figure 8. Bivariate map of Chicago ........................................................................................... 60
Figure 9. Bivariate map of Chicago at the census block geography ............................................ 66
Figure 10. Demographic profile of conventional high density census blocks ............................ 68
Figure 11. Demographic profile of green census blocks ........................................................... 69
Figure 12. Demographic profile of purple census blocks ......................................................... 71
Figure 13. Demographic profile of conventional low density census blocks.............................. 72
Figure 14. Demographic profile of orange census blocks ......................................................... 73
Figure 15. Demographic profile of pink census blocks ............................................................ 75
Figure 16. Poisson distribution .................................................................................................. 77
Figure 17. Individual representation map with image, Lakeview ............................................... 79
Figure 18. Individual representation map without image, Lakeview .......................................... 80
Figure 19. Individual representation map with image, Forest Glen ............................................ 81
Figure 20. Individual representation map without image, Forest Glen........................................ 82
Figure 21. Individual representation map with image, South Lawndale ..................................... 83

vii

Figure 22. Individual representation map without image, South Lawndale ................................ 84
Figure 23. Individual representation map with image, South Deering ........................................ 85
Figure 24. Individual representation map without image, South Deering ................................... 86
Figure 25. Vertical residential environment in Chicago ............................................................. 89
Figure 26. Mix use building, Lincoln Park Chicago ................................................................... 90
Figure 27. Conventional high density comparison ..................................................................... 94
Figure 28. High density, high personal space apartment ............................................................ 96
Figure 29. High density, low personal space apartment ............................................................. 96
Figure 30. High density, high personal space house ................................................................... 97
Figure 31. High density, low personal space house .................................................................... 97
Figure 32. Conventional low density comparison ...................................................................... 99
Figure 33. Low density, high personal space house ................................................................. 102
Figure 34. Low density, low personal space house .................................................................. 102
Figure 35. Low density, high personal space loft ..................................................................... 103
Figure 36. Low density, low personal space house .................................................................. 103
Figure 37. Individual representation map of Hyde Park, with image ........................................ 105
Figure 38. Individual representation map of Hyde Park, without image ................................... 106
Figure 39. Cabrini-Green redevelopment ................................................................................. 115

viii

Chapter 1
INTRODUCTION

1.1 Introduction
Contemporary cities represent a unique process of how populations settle on the earth’s
surface. According to the latest estimations, over half of the world’s populations now live in
cities. The percentage of people inhabiting the urban space is expected to grow in the future
(United Nations, 2010). Furthermore, the manner in which people settle within the landscape of
the city is changing. Residential patterns are transforming from suburban sprawl towards a more
centralized downtown pattern of residential structures (Kern, 2007). Increasingly, a higher
proportion of the world’s populations are going to be living in the vertical urban residential
spaces. Conventional metrics to measure and analyze populations capture the urban space
without taking into consideration widespread vertical living spaces. As a result, the current
conceptions of crowdedness and spaciousness in cities are inadequate.
Conventional density metrics characterize the population occupying a given areal extent,
such as people per square mile or people per square kilometer. This approach dates back to the
emergence of data driven thematic maps in the 18th century (MacEachren, 1979). Conceptions of
space in the early thmatic maps show a world where nearly all people lived on the earth’s
surface. One of the perhaps unintended results of the conventional density measurements is the
perception of crowdedness or spaciousness in the lived experience of people within the mapped
area. Conventional planar approaches measuring population spaces fail to capture the settlement
phenomena of crowdedness and spaciousness in contemporary vertical urban environments.
Conventional population density is a poor measure in contemporary cities for two distinct
reasons: 1) the assumption of homogeneity of individuals within the boundary space (Spielman

1

and Yoo, 2008) and 2) the assumption of homogeneity of the vertical spatial extents of
residential structures across the boundary space. Cartographic representations of cities often
show zones of singular values abruptly ending at defined borders with a choropleth design.
Consequently, conventional density and the subsequent cartographic products fail to capture the
reality of contemporary urban living. A new measurement is needed, a measurement that can
accurately calculate the amount of space each individual person has within the vertical urban
landscape of the city. This thesis introduces the measurement of personal space. The personal
space metric is aimed at combating the vertical space problem and providing a measurement that
matches the form and complexity of the contemporary city.
The main limitation of conventional population density in contemporary cities can be
illustrated with a quick example. Imagine two hypothetical settlements of one hundred people
living in a building with the same size footprint. In the first settlement, one hundred people are
distributed in a ten story building of fifty apartments with an average of 1,000 ft 2 of living space
per apartment- a total of 50,000ft 2 of livable space within the building. In the second settlement,
the same one hundred people live in a single story building with fifty apartments with a total of
5,000ft2. In the second scenario, each apartment in building has an average living space of
100ft2. The two hypothetical buildings illustrate how the metric of personal space, how the lived
experience of crowdedness or spaciousness within the residential structure is quite different
when the vertical extent of the buildings are accounted for.
The main problem with a conventional population density approach is the two drastically
different living spaces described above are characterized identically as homogenously distributed
population units. The metric of personal space allows for the obvious variations in living spaces
between the two buildings to be accurately characterized. A move towards the metric of personal

2

space rather than conventional density as the input variable in urban models could significantly
change the perception of urban space and allow for a more realistic distribution of people within
the urban environment to be understood and represented with more accuracy and precision.
Urban planners commonly use the measurements of Floor Area Ratio (FAR) and Floor
Space Index (FSI) to account for and regulate spaces within building, indicating a conscious
effort to account for spaces within individual buildings in the planning and operations of the
cities (Taylor, 1998). FAR and FSI are measurements which allocate how much floor space a
building can contain based on the size of the property and are included in zoning codes and laws.
The conceptualization of urban spaces by planners and civil engineers fails to translate to the
geographic and spatial analytic depictions of the city space. Urban planners account for the
interior space of buildings in the construction, modifications, and improvements to the city.
Traditional two dimensional metrics utilized by geographer are unable to differentiate the planar
surface boundary spaces with the vertical built spaces. Urban population maps fail to
conceptualize the three dimensional living space carefully constructed by urban planners.
Several prominent urban settlement models rely on conventional population density
metrics as a means to conceptualize residential concentrations. The Burgess Model, for example,
as with many others from the Chicago School, use population density as a parameter in the
conceptual models describing the structure, forms, and design of the urban space (Keating,
1998). Conventional population density has been instrumental in the development and
understanding of urban spaces across the world. Urban models of growth and form- Burgess’
concentric zone model, Hoyt’s wedges of activity model, and Harris and Ullman’s multiple
nuclei model, for example- all utilize the measurement of population density and the distribution
pattern of population densities as a critical input variable (Macionis& Parrillo, 1998).

3

Historically, urban models are created with the assumption that population density is an accurate
and explicit measurement of population patterns. The argument set forth is that conventional
density does not reveal the entire narrative about the urban landscape.
Population density is a 20th century measurement that is improperly used to characterize
21st century urban spaces. As urbanization rates rise throughout the world, more people will be
settling into the vertical urban environments. There is a critical need to improve population
metrics to account for the three dimensional residential space. Geographers can enhance the
understanding of 21st century urban studies by expanding the methodologies for quantifying the
distributions and densities of people in vertical urban spaces. The residential environment of
places like New York, Chicago, and San Francisco are characterized by the multistory residential
structures that can contain many hundreds of individual people in a singular surface space. The
contemporary urban residential environment includes an assortment of vertical living extents. To
understand how populations live in cities and to understand the perception of space and
crowdedness that conventional density ultimately measures, the metric itself must be rethought
and reworked.
Rethinking the methods in mapping of urban populations, deconstructing the boundaries
of neighborhoods and census divisions, and modeling the vertical living space are important
steps in creating a consistent measure to understand spatial structure of people in the modern
urban space.

4

1.2 Research Goals
The two primary goals of this thesis are:
1) To identify and examine the limitations of current methods of examining urban
population distributions, and
2) To design and implement new approaches to map and understand populations in the
modern urban environment.
I will approach this problem from three perspectives. The first approach will be to
understand the theoretical framework behind the construction of urban boundaries and how
these shape population analysis in the city of Chicago, Illinois, USA. This perspective will
look at the historical process in which the Chicago city boundary system was created and
identify the uses and limitations of this system in the current work in urban population
analysis. The second approach will be the creation of new models that deconstruct the
boundary space and reconstruct urban spaces with an emphasis on accounting for the vertical
spaces. The third approach is the creation new cartographic representations of the urban
space and explanations on how these new maps work to eliminate many of the errors and
misrepresentations of traditional techniques.

5

1.3 Research Questions
The thesis aims to answer three core research questions:
1. What are the limitations of traditional population density?
2. How do these limitations confuse our understanding of contemporary urban settlements?
3. How can emerging geographic analytics help address these limitations and generate a
more complete and precise understanding of urban population distributions?

This thesis aims to provide an alternative technique to understand and visualize urban
population distributions and the structure of contemporary cities. This thesis will explore a
variety of analytical techniques to create a series of visualizations that reveal a new interpretation
of populations in the urban space without the reliance on the established administrative boundary
system, the conventional analytical techniques, and the standard representations.

6

Chapter 2
BACKGROUND
2.1 Introduction
This section explores the study of urban settlement distributions and the cartographic
representation of human populations. It also examines the role, the origin, and the significance of
United States Census to understand the process in which population data is acquired and used in
research projects. This section also explores common population mapping strategies such as the
choropleth and dot density techniques. This section will also explore the cartographic products
derived from conventional population density metrics, looking at both common uses and
limitations of such outputs. Examples of how conventional population density has played a role
in the modeling and the constructions of the urban space are discussed. The function of
enumeration units and boundaries is discussed in the context of the urban space. The final section
introduces the theory of dasymetric mapping and discusses applications and limitations of this
approach in contemporary urban population mapping projects. The purpose of this section is to
explain where population data comes from, how it is commonly used in cartography, and how
emerging techniques can be employed to maximize the functionality of the data.

2.2 United States Census
2.2.1 Census Background
The utilization of the census as a measure to understand population has a long history in
Euro-American culture. The census was designed originally as a metric to measure military
strength, physical territory, and tax liabilities of colonial European powers. The modern form of
the census, where every individual within a defined political boundary is counted, began in the
colonial territories of western European countries in the eighteenth century (Anderson, 2000).
7

The United States Census Bureau conducts a census to determine how many people live
throughout the country and within defined smaller political boundaries such as states or counties.
A census is a count of everything within said boundaries at a given time. The United States
Census operates in four distinct parts working together to count every individual. First, a national
address list is established to account for all residential locations. Population information is then
collected for all of the address locations in the national database. Once completed, a follow-up
procedure is enacted to account for those who did not respond to the initial request. Finally, an
assessment model attempts to account for those not recorded in the first three steps. Despite the
intention to count all members of a population, the census suffers from two primary sources of
error- omissions and erroneous enumerations.
An error assessment performed on the 1990 United States Census show approximately 10
million errors of omission and 15 million errors of erroneous enumerations. On the national
level, this represents a gross error of a 5 million person undercount. At the census block level of
aggregation, the lowest level of census measurement, there is a gross error of 25 million people
that may be falsely assigned or miscounted (Anderson, 2000). Despite these errors, the census is
regarded as a highly accurate and scientific collection of population data within multiple
boundary enumerations.
There are multiple enumeration units which the United Sates Census data are aggregated
to and released to the public. There is a geographical hierarchy to the geographies of the census
data structure. The hierarchal order, moving from largest to smallest units, consists of the
following geographies: nation, state, county, county subdivision, place, census tract, census
block group, census block (United States Census Bureau, 2010). The term county subdivision
refers to cities and municipalities within a county; place refers to neighborhood units within a

8

city. Census blocks are the smallest geographic unit used by the United States Census Bureau
and was first used in 1940. There are 8,200,000 census blocks in the United States. A census
block group is a continuous cluster of census blocks. Block groups are designed to have roughly
1,500 people and can vary in size and number of individual blocks based on the concentration of
people in the vicinity. Census tracts are typically small and relatively stable units. The optimal
size of a census tract is 4,000 people. As a result, census tracts in urban spaces are
geographically smaller than census tracts in rural spaces (United States Census Bureau, 2010).
A function of a census is to assess the health of a population- to understand if the total
population for a given area is growing, declining, or maintaining at a stable rate. The primary
intention of the United States Census is to collect the spatial locations of the population to assign
equal and fair representation in the congressional sector of the government. A census also
provides demographic characteristics about a population; for example, the census records race,
age, sex, marital status, and income level of an individual. The goal of a census is to understand
the characteristics of the residents of a society (Langford, 2003). Characteristics of population
can include race or ethnic self-identification, household income, household size, and land use
characteristics of residence.
The United States Census began in 1790 and has been conducted decennially since
(Laven, 1996). The first census law, ACT 1 of 1790, establishes the ‘residence rule’- instituting
the policy that a person is to be counted where he or she resides (Anderson, 2000). This rule
establishes the idea that the goal of the census is to understand where people live and how they
should be taxed and represented accordingly. The goal of the census is not to determine the
movements or actions or people in a given day but rather to count how many people reside in
each of the designated political boundaries. Data collection by the United States Census Bureau

9

consists of two questionnaires: a short form and a long form. The short form is limited in the
number of questions and is asked to every individual or housing unit. The long form is a sample
of population characteristics, asked of roughly one in six households, and contains questions
more focused on socioeconomic and migration information (United States Census Bureau,
2010). Census results are compiled and released in the following formats:
1.Summary File 1 (SF1) - General information on U.S. population distributed down to the
smallest geographic unit, census block level.
2.Summary File 2 (SF2) - General information on U.S. population focused on race and
ethnicity measures. Distributed at the census tract geographic level.
3.Summary File 3 (SF3) - Largest database released by Census Bureau. Contains many
attributes concerning race, ethnicity, housing, and income and is distributed at the block
group level.
4.Summary File 4 (SF4) - Information derived from the long form sample of population
characteristics. Distributed at the census tract level.

2.2.2 Importance of Census
The census is one of the cornerstones of the United States democratic government.
Representation in the congressional sector of government is determined by the population of
each state as counted by the census every ten years. Reallocation of congressional representation
is performed after the population data has been calculated from each census. The differences
among a state’s population, land mass, and natural resources is a method to help determine each
states financial contribution to the federal government. Population size is one of the measures of
each states representation within and financial support from the federal government. Census data

10

is important beyond the original intentions; the data provided by the United States Census has
become useful outside the federal government and is often used by planners, resource managers,
and academic researchers (Lavin, 1996).

2.3 Population Density
2.3.1 Population Density Overview
Conventional population density measures the concentration of people within a
standardized geographic unit of measurement. This measurement traditionally involves units of
ground area, square miles for example, to provide a quantification of the amount of people per
unit of space. A comparison of population concentrations across varying size geographies is a
key output of the density measurement. Utilizing population density allows for comparisons
across geographies. Large urban areas, city-states, microstates, and dependencies are often the
most densely populated regions in the world (United Nations, 2010) with relatively small
geographic areas and an economically specialized population drawing on the nearby resources
from rural surroundings. High density spaces are most often clustered around seaports,
transportation hubs, and areas with high amounts of fresh water (Shah, 2002).

2.3.2 Alternative Density Measurements
Conventional population density measures the number of people within a defined
geographic areal unit. It is a standardized measurement that can compare rates of population
across geographies. For example, Texas and Rhode Island can be compared without the size of
the two states geographic areas influencing the results. Population density eliminating the errors
or misrepresentations attributed to the size of the enumeration unit- the modifiable areal unit

11

problem (MAUP). Several other methods of calculating population density that are specialized to
particular fields try to provide more accurate measurements based on land cover and land use
characteristics. The most common form of population density is arithmetic density, a simple rate
of the number of people per area of land. Arithmetic density does not take into account other
factors such as land use patterns or demographic attributes (Trewartha, 1953).
Physiological density measures the number of people per unit of arable land within a
larger boundary. An example of this would be all the people in a state divided by the amount of
land suitable for growing food within that state. This measurement is aimed at quantifying the
food security of a population. Similarly, the measurement of agricultural density measures the
number of people per area of cultivated land (Brown and Podolefsky, 1976). A related
measurement is ecological optimum density. This measurement calculates the maximum density
of population which can be supported by the natural resources of the enumeration unit. This
measurement uses an array of land use, agricultural, and socioeconomic ancillary data to
determine natural resource wealth and how many resources an individual requires for survival.
A pair of related density measures, residential density and urban density, aim to explain
the population environments of urban spaces (Beckman, 1969). Urban density measures the
number of people inhabiting a unit of urban land. Urban land is defined as places classified urban
with traditional land use/ land cover analysis. Urban land cover includes residential, commercial,
industrial, and open spaces within the greater urban landscape. Residential density measures the
number of people living per unit of residential space. Residential space is defined by an urban
space dominated with housing units. Housing units include single family, multi-family, and
mobile homes of varying heights (Griffith, 1981). Residential density begins to measure
populations in units of livable space but continues to utilize planar surfaces in analysis.

12

Conceptually, the idea of overpopulation is often linked with high population density.
Spaces are considered to be overpopulated when the number of humans living in an area exceeds
the carrying capacity of that area and does not necessarily correlate to high population density.
The carrying capacity of an urban space is dependent on quality of housing and infrastructure,
access to resources, and ability to move across the spaces of the city (Aina, 1992). There is
debate of whether administratively autonomous suburban communities that surround and support
the city should be accounted for or if solely the area within the city boundary be used in urban
population analysis (Macionis& Parrillo, 1998).
A metropolitan area includes the city as well as the satellite and suburban areas around
the city. A metropolitan region can include nearby rural and farm land that is socio-economically
connected to the city. Determining the appropriate extent of metropolitan Chicago can vary
depending on how the characteristics of the metropolitan space are defined. It is difficult to
quantify how much of the resources of Chicago are dependent on the suburbs and vice versa. It is
clear that the daytime population of Chicago and the residential living experience of the city are
affected by the suburban pattern of development, but the extent of this influence has not been
investigated for the purpose of this thesis. The population density of Chicago will be
significantly lower when measuring the entire metropolitan region as opposed to just the city
itself due to the suburban sprawl and rural spaces just outside the city limits (Macionis &
Parrillo, 1998).

2.3.3 Population Density Limitations
The primary limitation of the conventional population density measurement is that it
represents the number of people per two-dimensional space. Two-dimensional analysis does not

13

reflect the manner in which people are distributed across the urban space. People are not
distributed solely across a flat surface, especially in contemporary urban spaces. Apartments,
condos, and houses create a three dimensional environment conventional population density fails
to account for. Due to this limitation, conventional density measurements overestimate the
crowdedness of areas with high vertical extents. In vertical residential environments, many
people live within the planar surface of the boundary but can be distributed into an array of
different vertical spaces. Consequently, the conceptualization of crowdedness and spaciousness
may be skewed. Conceptual models of urban environments using conventional population
density as an input variable are outdated for the contemporary city.

2.3.4 Population Density in Models and Applications
Urban theories use conceptual models to characterize the dynamic processes of
expansion, growth, reclamation, and movement that give a city its character and form (Burgess,
1925). Population density is at the core of urban models and is often used as an input variable in
the development of theories aimed at understanding complex patterns of urban life. The most
famous of these theories is the concentric zone theory developed in 1924 by Ernest Burgess, a
sociologist at the University of Chicago (Macionis& Parrillo, 1998).
Using the urban framework of Chicago as a study area, the concentric zone model
explains the spatial distribution of social groups within the urban space. The concentric ring
model depicts urban land use in concentric zones away from the city center. The Central
Business District (CBD) is in the middle with concentric rings of lower intensity and lower
density land uses expanding outward towards the suburbs. The zones are characterized by the
land use patterns, conventional population densities, and socioeconomic attributes in linear

14

buffers away from the CBD. Outward from the CBD are the transition zone, the working class
zone, the suburbs, and the commuter zone (LeGates and Stout, 1996). The Burgess model is
based on the urban landscape of 1920’s Chicago which did not have the same vertical
environment as the contemporary space.
Expanding on the concentric zone model and also using Chicago as a social laboratory,
Homer Hoyt developed the sector model in 1939. The sector model allows for an outward
progression of growth along transportation networks rather than by Euclidian distances.
Beginning in the CBD, the zones expand outward from the city center along railroads, highways,
and other transportation arteries. The sector model characterizes the evolution of upper class
residential sectors evolving in the North Side of Chicago along Lakeshore Drive, middle and
lower class development patterns along the elevated rail lines to the west, and industry extending
southward along heavy rail lines (Hoyt, 1939). The inclusion of transportation routes allows for a
model with the input of mobility added to conventional population density and socioeconomic
characteristics previously observed by Burgess. The resulting pattern of development is
characterized by wedge-shaped sectors expanding outward from the CBD along different types
of transportation corridors (LeGates and Stout, 1996).
The concentric zone model and the sector model are two examples of how conventional
population density can be utilized as an input in development theories. Other notable theories
using conventional population density as an input metric include Harris and Ullman’s ‘multinuclei model’ (1945) and Soja’s ‘polycentric archipelago’ analysis in Los Angeles, California
(1989). The significance of these urban models in laying the groundwork towards a greater
understanding of urban spaces is not to be dismissed. Moving forward, however, to create
accurate models of rapidly changing urban spaces, the need for highly accurate input variables is

15

essential. Conventional population density is inappropriate for quantifying the crowdedness or
spaciousness of contemporary cities with expansive vertical urban spaces.

2.4 Cartographic Representation
2.4.1 Introduction
Cartographic representations of population data often employs a thematic design to
provide specific information about the nature of the population distribution and to allow for the
comparison of spatial patterns between geographies. Population concentrations can be mapped
with a variety of techniques, all with certain advantages and limitations. In this section, I will
focus on two types of cartographic techniques for representing population- the choropleth and
the dot density design. I will focus specifically on how these approaches are misused and, in
some cases, inadequate for precise map designs of population distribution. The concept of
dasymetric mapping is introduced and consideration is given to the ways this technique can
efficiently create more intelligent and accurate maps of population distributions.

2.4.2 Choropleth Maps
Choropleth maps are a type of thematic map that show the distribution of observations
based on different geographic units defined by boundaries. Dasymetric and choropleth maps
have a similar purpose and origin, but over the past century the choropleth map has become the
more popular cartographic output (Eicher & Brewer, 2001). Choropleth mapping is an
appropriate technique for representing a phenomenon that is uniformly distributed across the
geographic extent. This method is most effective when the geographic units being mapped are
similar in size (Slocum et al., 2009). Similar size geographies will limit the influence the MAUP

16

has on the data and the error introduced when geographic data is aggregated to spatial units for
analysis.
The MAUP and the concept of ecological fallacy address two main concerns with
geographic data representation (Openshaw, 1984). The first is the problem of scale when the
areal unit is modified. This geographic uncertainty occurs when data from one unit are
aggregated into different size units of analysis. For example, the aggregation of state data to
county data will dramatically change the results and the distributions of the observations and
introduce a degree of uncertainty to the analysis. The aggregation error also occurs when the
analysis is performed across differing geographic units. For example, errors will occur when
comparing state and county data in a singular analysis.
Ecological fallacy is closely related to the MAUP and occurs when results based on the
aggregated data are used to describe the individual observations within the zones or groups being
studied. All individuals within a county do not share a common attribute; the choropleth
technique, however, classifies the county as a singular value and represents that with a singular
visual variable. Election maps illustrate this problem- when 60% of a county vote is for the
Democrat canidate, then the county is represented as blue on a choropleth election map. This
visual representation, however, does not speak to the party affiliation of all citizens of the
county. Choropleth maps are generally used to visualize quantitative data with discrete
boundaries but are subject to these limitations brought on by the MAUP and the theory of
ecological fallacy.

17

2.4.3 Choropleth Maps and Population Density
Population density is often represented with choropleth maps that visualize the rate of
people per area of ground unit with a singular visual variable within a polygonal enumeration
unit. Population density, in the choropleth design, is best represented with the visual variable
lightness, showing sequential differences between different geographic units (Langford &
Unwin, 1994). A choropleth map assigns each polygon a single attribute that changes abruptly at
the border. The abruptly changing nature of the choropleth map is not necessarily an appropriate
representation for the continuous distribution of population concentrations (Holt et al., 2004).
The distribution of people in space can be conceptualized as a continuous phenomenon, a
geographic event that occurs everywhere rather than an event that occurs within a set of
boundaries.
The choropleth mapping technique breaks the field of population data into classes with
differences defined at administrative boundaries that may or may not reflect population clusters
and patterns. Modifying the areal units by changing the type of geographic boundary used in the
map can greatly change the results of the analysis (Openshaw, 1984). A neighborhood could
show an above average population density, but when aggregated to census blocks, the results
may show a wide range of densities. Additionally, the choropleth technique also gives the
impression of abrupt changes between boundaries of geographic units when change is often more
gradual. The boundaries of a neighborhood or a census tract are often arbitrary to the settlement
patterns and do not reflect breaks in the distribution of a population (Brewer & Eicher, 2001).
A second issue with the utilization of choropleth maps for population density is the
impression that the population is distributed evenly or homogeneously across the extent of the
polygon unit (Maantay et al., 2007). In reality, populations are often clustered, stacked on top of

18

each other in buildings, and coincide with the patterns of development within the geographic
units used in the map. Human populations are not distributed uniformly across the geographic
space; the choropleth technique gives the direct impression they are.

2.4.4 Dot Density Maps
Dot density maps use dots to represent concentrations of a feature across a spatial extent.
The population of an entity determines how many dots are placed within a specific polygon
boundary. For example, if a county polygon has a population of 10,000 people with one dot
equaling 50 people, there would be 200 dots of the same size randomly located within the county
boundary. The dots in dot density maps do not represent the specific location of an entity but
rather a summarized collection of individual observations within the boundary of an area
(Slocum, T. A. et al., 2009).
A major limitation of the dot density visualization is, due to the random placement of the
dots within the geographic boundary, dots are often located in places where it is physically
impossible for a population to be. For example, a dot density map of human populations
randomly placing one dot for each 500 people within county may result in the placement of dots
into water, open spaces, industrial areas, and commercial zones (Goldsberry, 2010). When using
a dot density technique, it is optimal to have the smallest spatial units possible and to incorporate
ancillary data to remove unsuitable locations. This allows for the placement of the dots to be
more precise in geographic location and show more accurate distributions of the population.
A second limitation of the dot density technique is the value assigned to each dot creates
what could be an unnatural clustering of observations and could be misleading the map reader.
An example of this problem is most obvious in designs where a secondary visual variable such as

19

hue, lightness, or saturation of the dot represents an attribute about the population. If the dot is
representing race, for example, like races would be clustered together within the boundary of the
polygon. Referring back to the hypothetical county with a population of 10,000 people
represented by 200 dots, each dot representing 50 people, this issue becomes problematic and
misleading. Every dot will show a clustering of 50 people of a like race, a representation which
will greatly exaggerate the racial segregation of the county. The dot density approach, if done
without proper consideration, can imply racial clustering that may or may not exist and even the
most diversely populated, integrated spatial units could appear quite segregated (Goldsberry,
2010).

2.4.5 Dot Density Considerations
Four main considerations must be carefully managed when making a dot density map in
order to intelligently design the map while limiting the amount of misleading variables. The
cartographer must consider the aggregated boundary units, the size of the dots, the observations
per dot, and the location of the dots.
When making a dot density maps, the smaller the polygon the dots are randomly placed
within the more accurate the location will be to the entity being mapped. Ancillary information,
such as land use, land cover, and zoning information, can be used to choose the correct size and
type of geographic boundaries to place dots within. The dot size used in a dot density map can
affect ability of the map reader to perceive the density of the phenomena accurately. If the dots
are too small, the resulting pattern will show a sparse density. If the dots are too large, the
represented population can appear overly dense (Slocum, T. A. et al., 2009). Another important
consideration in how many entities in the population should be represented by each dot- the unit

20

value. If the unit value is too small, the resulting map will show too many dots in each
geographic unit and make the phenomena appear denser and more clustered than it is in reality.
If too large of a unit value is assigned to each dot, the geographic unit will appear too sparse to
accurately show the intended spatial relationship of the data. Finally, one has to determine the
precision of location for the dots in the map. If the purpose of the map is to show the absolute
location of the population within a greater boundary, a dot density approach would be
inappropriate (Slocum, T. A., et al., 2009).
The main advantage of dot density maps is the ability to represent a phenomenon that
transforms gradually over geographic space. The relative crowdedness or sparseness of the dots
displayed on a map will visually match the concentrations of the population. Dot density maps
also have disadvantages in representing the spatial distributions of populations. Due to the
random placement of the dots, there may be no actual events of the phenomenon that is being
mapped where the dots are physically placed on the map. Additionally, the subjective nature of
the dot and unit size decisions can skew and distort the relative density of the phenomena being
mapped.

2.4.6 Dasymetric Maps
Dasymetric mapping is a method that can represent the population distributions and
densities more accurately than either conventional dot density or choropleth techniques by
incorporating ancillary information. Dasymetric maps reflect the population of an area with
greater accuracy by allocating specific areas of suitability within an enumerated areal unit
(Brewer & Eicher, 2001). Choropleth maps use existing administrative boundaries that are
frequently independent of the phenomenon being mapped. In dasymetric mapping, ancillary data

21

divides the current larger administrative boundaries into smaller spatial units of suitability (Holt
et al., 2004). For example, the population of a census tract in a choropleth map will show an
evenly distributed population across the entire unit space. The dasymetric method separates the
residential areas from the commercial, natural, or industrial areas within the census tract and
allocates the total population to only the residential sections. The dasymetric method works to
eliminate the ecological fallacy inherent to the choropleth technique, making for more accurate
representations of the distribution of people. Creating maps that accurately represent the density
and distribution of people is essential for providing services and assessing risk to urban
populations (Maantay et al., 2007).

2.4.7 Dasymetric Methods
Dasymetric mapping allows for the allocation of populations into smaller units with the
utilization of ancillary data. Previous studies have shown this to be an effective method to
represent population distribution with greater accuracy. Refining the units to reflect residential
suitability typically uses land use/ land cover data ancillary information. John K. Wright first
used dasymetric maps in a population analysis of Cape Cod in 1936. His study redistributed the
population of the area into inhabited and uninhabited regions as indicated on the United States
Geologic Survey (USGS) topographic maps. He then subdivided the inhabited areas into smaller
units based on various settlement patterns and a controlled guesswork approach from inherent
knowledge of the region. The theoretical concepts employed by Wright in early dasymetric
approaches are still in use today, commonly with ancillary data derived from remotely sensed
images (Liu et al., 2008).

22

Land cover classifications derived from remotely sensed images provide accurate data on
places of human populations. Advancements and higher resolution images have led to greater
accuracy in the separation of the residential and non-residential spaces in the urban environment.
Detailed ancillary data, an understanding of urban population tendencies, and knowledge of the
study area are all important factors in creating dasymetric zones for accurate population mapping
(Brewer & Eicher, 2001). There are many dasymetric methods, each with certain advantages and
disadvantages, to utilize ancillary information and allocate population data into smaller units.

2.4.8 Binary Method
The binary method is the simplest method in dasymetric mapping, using the ancillary
land use/ land cover data to exclude areas where populations are not present. This method takes
the aggregated population data within the target zones and eliminates the non-residential areas,
redistributing the populations into the areas that meet the condition of the binary statement
(Maantay et al., 2007). The binary method in the context of urban data deems single-family and
multi-family housing areas as suitable for residential and places of business, industry, and natural
areas as unsuitable. This method is often the first step in the dasymetric process to filter or mask
out the areas deemed unsuitable (Brewer & Eicher, 2001). The creation of dasymetric maps with
the binary method eliminates the generalizations resulting from the utilization of arbitrary
boundaries in the choropleth technique.
The binary method suffers from limitations in assuming homogenous spaces across all
residential areas of the urban environment. Cities have many different residential structures, from
high-rise apartments to single story houses, all of which have vastly different population
distributions and densities. Not all parts of a city have residential pattern of development that can

23

be accurate characterized with a simple binary expression. With the binary method however,
every zone meeting the requirements of the binary system are considered suitable for the
residential classification (Mantaay et al., 2007). Additionally, a census block that is a majority
non-residential could fail to meet the requirements of the binary method, even though a portion
of the census block may contain a residential sector. The population within this unit fails to meet
the requirements of residential suitability and is not represented on the map appropriately.

2.4.9 Areal Interpolation Method
The areal interpolation method allows for the transformation of a source data set into a
target data set (Mennis 2003). This method is used to aggregate population values for the
different zones within the target layer. Within the context of dasymetric population mapping, the
source data are the census population data and the target data are the land use/ land cover layer
(Liu et al., 2008). The areal interpolation method will group data within the specified zones from
the source data and aggregate it into the appropriate zones of the target data set. Areal weighting
is a simple interpolation method which allocates the summarized population data according to
the proportional areas of the zones in the target data (Langford, 2003).
Areal weighting operates on the assumption that populations are distributed uniformly
within the target zones (Maantay, 2007). A weighted areal interpolation scheme has been used to
allocate each county data set on population, income, and employment into hydrological basin
target zones in California to gauge the impact of socioeconomic variability on water usage and
public policy (Goodchild et al., 1993). This study operates on the assumption that the population
density of the counties is uniform when moving from the fifty-eight observations of source data
to the twelve observations in the target data. This method proves useful in allocating data from a

24

source data set to a target data set and showing a population distribution within weighted zones;
the assumption of density contingency, however, leads to high amounts of uncertainty.
Areal interpolation is an approach that applies directly to dasymetric mapping of
population density but has limitations by representing populations or observations as evenly
distributed events across the target zones. Although the areal interpolation method reflects a
more accurate geography, the purpose of dasymetric maps is the move away from the
homogeneity expressed by the choropleth map design (Holt et al., 2004).

2.4.10 Three Class Method
The three class method assigns a weighted value to different land use classes to represent
clusters of populations in certain types of environments. An example used to map population
distributions in rural Montana allocated 80% of the population to urban zones, 10% of the
population to open zones, and 5% each to agriculture and forest zones (Holloway et al., 1996).
The assumption in the model was that a majority of people would live in the urban zone while
scattering of individuals would occupy the traditionally non-residential zones. A second study,
attempting to replicate the population patterns in relation to land use, employed a three class
method to allocate 70% of the population to urban, 20% to agricultural, and 10% to forest in
Pennsylvania (Brewer & Eicher, 2001). The three class method is used to distinguish land use
zones with a certain percentage of the population and then to allocate individuals homogenously
across those zones. The three class method assumes even distribution of populations across the
zones and does not account for variations in urban development. Another weakness of this
method is that the process assigning the percentages to each zone is subjective and may or may
not be based on demographic evidence or personal knowledge (Maantay et al., 2007).

25

The three class method is also subject to the modifiable areal unit problem as a study area
may have a very small urban zone and a very large agricultural zone. In this scenario, 70% of the
population would be distributed in the small urban zone and only 20% across the vast
agricultural zone. The distribution of the population reflects an error attributed to the sizes of the
zones. Furthermore, although this method recognizes differences between land use classes, it
does not account for differences within the zones. Population distribution is not constant across
all urban land use zones (Liu et al., 2006). Remotely sensed images provide little insight into the
variability within urban zones- differences between types of residential environments. Ancillary
information such as zoning designations, ownership maps, or municipal data sets is the most
appropriate data for the differentiation of urban space. The three class method has a great deal of
potential in creating accurate dasymetric maps, especially at larger scales where issues of
variability within urban zones is less of an issue.

2.4.11 Limiting Variable Method
The limiting variable method is an extension of the three-class method, using the same
general theory but setting a maximum threshold on the amount of individuals that can occupy
each polygon zone (Maantay et al., 2007). This is done by setting a maximum density to each
land zone determined to be inhabitable. When the population reaches the allocated threshold of
maximum density, observations will be systematically distrusted to other acceptable zones at
proportional rates (Brewer & Eicher, 2001). This method is commonly used to ensure the
agricultural and forested zones on the map are not over populated under the assumption that
these areas will be sparsely populated. When the allocation to agriculture reaches the maximum
density, the remaining observations are transferred to another zone, typically the urban zone

26

(Slocum et al., 2009). The limiting variable method works to eliminate the MAUP that exist
between substantially different size zones and maintains assumed population patterns. The
limiting variable method also assumes populations are equally distributed across the zones as
well as introducing a subjective cutoff point that may or may not reflect ground conditions. Of
the methods studied, the limited variable method shows the lowest statistical occurrence of errors
(Brewer & Eicher, 2001).

2.4.12 Limitations in Dasymetric Mapping
Dasymetric mapping provides a better estimation of population distribution and density
than conventional choropleth and dot density methods. Despite the improvement over
conventional techniques, systematic flaws persist that make it a less than perfect solution
(Langford & Unwin, 1994). By allocating population data based on ancillary land use
information, errors and uncertainties can propagate through the steps and negatively impact the
final output (Yuan et al., 1997). Inaccurate classifications of land use data can greatly alter the
accuracy of the dasymetric mapping process. The standard acceptable error for classification of
land cover/ land use data from remotely sensed images is 85% (Anderson et al., 1976). Remotely
sensed data are the primary source of ancillary data and an error of up to 15% creates a great deal
of uncertainty throughout the study.
There is also a subjective quality to dasymetric techniques. In weighted analysis,
assumptions of population patterns are often based on urban theories and the cartographer’s
knowledge of the study area. In each of the different methods commonly used for dasymetric
maps, there is an assumption that populations are distributed evenly within the designated areas,
that allocation rules are homogenous for all equal zones, and that population follows

27

mathematical models. Geospatial information is an abstraction of reality that model complex real
world phenomena rendering it useful and understandable. Uncertainty and error is present in all
forms of this abstraction (MacEachren, 1992). The dasymetric technique aims to reduce the
uncertainty of the geographic space with the allocation of population zones within larger
boundaries. The accuracy of the ancillary data used in this allocation process, however,
introduced a new set of uncertainty to the study.

2.4.13 Cartographic Representation Summary
There are many types of cartographic techniques to represent the phenomenon of
populations across a geographic space, each with certain advantages and limitations. The two
most common types of maps geographers create in population studies are choropleth and dot
density maps. Dasymetric techniques incorporate ancillary data to refine the geographical areas
and boundaries where individuals reside in an attempt to improve the accuracy of population
distribution maps. Despite the limitations, the dasymetric techniques provide a more in-depth
analytical tool and the opportunity to intelligently incorporate as much additional information as
possible about the geography of the place mapped. For the purpose of population mapping, a
dasymetric technique can eliminate areas where people do not live and aggregate the populations
into areas that more closely represent the geographic extent of living spaces.

2.5 Literature Review Summary
This chapter has reviewed the literature surrounding census information, the measure of
urban population density, urban models and applications using population density, and different
cartographic representations of this measurement. Choropleth and dot density maps are often

28

used to represent the spatial distribution of populations in cartography. Dasymetric mapping
improves upon the limitations of both techniques but also introduces a unique set of errors and
uncertainties with the inclusion of additional data. Dasymetric techniques rely on ancillary data
including land cover datasets, remotely sensed images, and zoning laws. Methods to combine
raw population data with ancillary datasets to create more accurate population maps include the
binary method, areal interpolation method, the three class method, and the limited variable
method. Dasymetric mapping has the potential to be a driving force behind a new urban
population mapping technique to account for and represent the three-dimensional urban space.

29

Chapter 3
STUDY AREA

3.1 Chicago Introduction
This thesis investigates the phenomena of population distribution in contemporary urban
spaces using Chicago, Illinois, as a case study. Chicago is the third largest city in the United
States with 2.7 million people in the city proper boundary (US Census Bureau, 2010). The city is
a unique urban space for the initial analysis both in terms of the historical construction of the
boundary system and the modern vertical form.
Chicago is considered the birthplace of the skyscraper. Louis Sullivan first recognized
that the skyscraper would represent a new form of architecture in the modern landscapes of cities
(Kaufman, 1969). Sullivan discarded the conventions of his time and designed buildings that
emphasized the vertical form before adequate technology existed to construct the design. Design
spawned materials and technologies in order to realize the Sullivan vision for the new urban
space. This new form of architecture, with an emphasis on the vertical structure, became known
as the Chicago School of Design (Kaufman, 1969).
Chicago boasts the tallest building in the United States, the Willis Tower, formerly Sears
Tower- a 110 story, 1,454 foot structure in downtown Chicago (Fountain, 2001). The history of
skyscrapers in Chicago began in 1885 with the completion of the Home Insurance Building, the
first steel-framed skyscraper in the world. Historically, Chicago has played a prominent role in
the development of the skyscraper and, at various times in the history, has featured the world’s
tallest building (Daniel & Grant, 2005).
Through the 20th century, Chicago went through two distinct phases of high-rise
construction- a first boom, from the early 1920s to the mid-1930s, and a second boom from the

30

early 1960s until the present. In both cases, high-rise structures were spatially concentrated in the
downtown urban core and in the semi-periphery immediately north. As of 2008, Chicago has a
total of 2,265 high-rise (over 150 feet) structures in the city; 903 of these, 40%, are zoned for
residential use. The structure of Chicago provides a unique study area for the creation of a new
methodology to model vertical urban living spaces.

3.2 Boundaries in Chicago
Boundaries play an essential role in our ability to understand the construction of space in
urban geographies. Boundaries represent the borders between the areal units that commonly
show differences in attributes across space. Boundaries clearly define the administrative pieces
of the urban space where structural characteristics such as identity, cultural capital, cultural
membership, commensuration, census categorization, political affiliation, racial and ethnic
distribution, and hegemonic power structures are organized and easily understood (Lamont and
Molnar, 2002).
In Chicago, the community area boundaries units are static delineations, unchanged since
their creation in the 1920’s. Movements of people, processes of redevelopment, immigration,
migration, and improvements or decay of the built environment have not changed the community
area boundary system. The community area boundaries of Chicago are shaped by the historical
context of the city’s growth, the annexation of suburban lands by the city in the late 19th century,
and the response to rapid urbanization by planners and urban theorists. The community area units
create unique spatial identities of the urban space and shape the context of the lived experience
for Chicago resident.

31

Cartographically, the community area boundaries create homogenous regions of a
singular attribute that abruptly change at a defined border. The utilization of the boundary units
for map making processes greatly simplifies and generalizes the complex and unique urban space
into a series of homogenous polygons. Conventional population density as defined by
administrative and generalized boundary demarcations is an unsuitable measurement for the
complex population processes within the contemporary urban space.

3.3 Growth of Chicago
“In his 1963 work, Victorian Cities, British urban historian Asa Briggs coined the term
‘shock city’ to describe cities that flourished as a result of the industrial revolution while at the
same time being afflicted by slums and seemingly unmanageable social and environmental
problems. Manchester was the prototypical shock city of England; Chicago of the United States
(Channick, 2008). Following settlement and establishment of the city in the early 19 th century,
Chicago became a place with transportation and industrial potential due to the geographical
proximity to the Great Lakes and Mississippi watersheds (Randall, 1999). Chicago was officially
founded in the 1830’s and grew rapidly as the idea of a internal transportation network became
paramount for nation building and the movement of resources from the interior of the country
towards the east coast.
In 1840, the population of Chicago was roughly 5,000 people. By the 1860’s, the
population had exceeded 100,000 and Chicago was emerging as the nation’s western
transportation hub with over thirty train lines entering the city. Simultaneously, Chicago was
becoming a dominant processing center for natural resource commodities from the Great Lakes

32

region to be transported back east (Cronon, 1992). By 1870, Chicago had grown to become the
nation's second largest city and one of the largest cities in the world (Cronon, 1992).
By 1870, Chicago had a population exceeding 300,000 and was beginning to extend
geographically into the immediate suburbs along streetcar lines. As Chicago grew, the emerging
pattern of settlement laid the framework from which the socioeconomic distributions existing
today can be observed. The division of residential and industrial sectors emerges with the
construction of the streetcar system. Neighborhoods of new immigrant populations settled in
ethnic enclaves such as Ukrainian Village, Wicker Park, Little Italy, Greek Town, and Noble
Square. By 1890, the population had risen to over 1,000,000 and many of the immediate suburbs,
places such as Hyde Park and Ravenswood, were annexed by the city. By the turn of the century,
the modern structure and the geographic extent of Chicago has formed during the period of rapid
urbanization and industrialization of the 19th century (Keating, 1998).
The introductions of the streetcar system had a significant spatial impact on the
settlement patterns of the new immigrants to the city. The neighborhoods of Hyde Park, Austin,
Norwood Park, and Ravenswood, each initially the end of a streetcar line and a set distance from
downtown Chicago, all experienced drastic population growth and building construction. The
space of Ravenswood, as described by Keating in Building Chicago, exemplifies this process.

A farming district in 1869, by the 1880’s it was one of the most exclusive suburbs in the
Chicago area and considered one of the finest suburbs in the country. Lawns were
spacious and well landscaped, houses set far back from the road and trees lined the
parkways. Most of the residents in Ravenswood by the turn of the century were
professionals and businessman who worked in downtown Chicago- lawyers, doctors,
dentist, heads of manufacturing companies, and real estate investors (Keating, 1998.
pp24).

33

The Great Chicago Fire of 1871 burned most of the existing infrastructure of the city,
destroying in excess of 18,000 buildings and leaving more than a third of the population
homeless (Pierce, 2007). Almost immediately, reform began in the city's construction and
building standards which allowed Chicago to emerge from the fire with a focused mindset on
constructing a modern urban space. The fire, in many ways, cleared the old and unwanted
infrastructure of the shock city expansion and allowed Chicago to start again from scratch with a
clear idea of the population growth, limitations of old construction techniques, and a focus on
modern design (Pierce, 2007). The fire thrust the city to the forefront of modern architecture.
Development and reconstruction of the city was created on a grid system to allow for more
simplicity in transportation and modern services, propelling Chicago into a status on par with
New York as the most important city spaces in the country by the early 20 th century (Randall,
1999).
Post fire construction and modernization allowed for the population in Chicago to
continue to rapidly grow in a planned and organized manner. By the turn of the century, the
population had risen to 1,600,000 and the elevated train, the ‘L’, had opened to allow for greater
growth away from the central city (Keating, 1998). The Loop, as the central city became known
as due to the pattern of the train network encircling the downtown area, became one of the
dominant central business districts in the country. Initially, the immediate zone around the Loop
consisted of many high density neighborhoods and ethnic enclaves consisting of mostly laborers
for the manufacturing focused downtown core. The ‘L’ allowed for greater access to the
suburban outskirts of the city and annexation of land continued supporting the continuing rise in
population (Pierce, 2007).

34

The residential growth boundaries in the Chicago area at the time were developed
primarily through the acquisition of rural land that was annexed by the city to provide modern
services the rural township governments were unable to provide for a growing suburban
population (Keating, 1998). The rapid rise in population and the annexation of the outlying
regions into the city of Chicago during this period created the structure of what would become
the community area boundary system.

3.4 Community Area Boundary System
In the 1925, The Social Science Research Committee at the University of Chicago
defined seventy-five community areas to separate the modern space of Chicago. At the time, the
measurement of community areas corresponded roughly to the ethnic neighborhood divisions,
physical boundaries, and the annexed suburbs acquired during the post-fire urbanization plan of
the late 19th century. Initially this process of community area subdivision was designed as a
management and resource allocation strategy; a mechanism to define the spaces and extents of
the neighborhoods that had grown in the initial industrial shock period and post fire
reconstruction of the city (Cronon, 1992). In the 1950’s, O’Hare was annexed and became the
76th community area. In the 1980’s, the community area of Uptown was split into two separate
community areas, Edgewater and Uptown. Other than these two additions, the boundaries have
not been revised to reflect change but instead have intentionally remained stable under the
premise that comparisons over time would provide valuable insight into the growth and health of
the urban space (Keating, 1998).
The community areas now divide Chicago into 77 community areas. These community
areas are well-defined within the consciousness of the Chicago residents and are static (Rankin,

35

2010). Census data are measured and represented at the level of the community areas.
Community areas also serve as the basis for a variety of planning decisions on both local and
regional levels including allocation of resources and collective funds. The measurement of
‘community area’ has replaced ‘neighborhood’ as the official static boundary designation for the
city as the term neighborhood has become a marketing tool of development and gentrification in
recent years (Keating, 1998).
Historically, the community area and neighborhood boundaries aligned; in the modern
space, the terms neighborhoods and community areas are no longer the same. Although many
community areas and neighborhoods share names and boundaries, many of the community areas
no longer correspond to any single neighborhood. There are over 200 designated neighborhoods
in Chicago, and unlike the larger, well-defined community areas, Chicago neighborhood names
and boundaries are notoriously unpredictable (Keating, 1998). In the modern Chicago
geography, community areas are divided into nine broad geographies: Central Chicago, North
Side, South Side, West Side, Far North, Northwest Side, Southwest Side, Far Southwest, and the
Far Southeast (Keating, 1998).
Proximity and direction from downtown Chicago determine the names of the nine larger
community area regions. Central Chicago consists of the community areas of The Loop, Near
North, and Near South and serves as the city's commercial and financial hub (Reiff, et al., 2005).
Central Chicago is home to the iconic city symbols of Grant Park, the iconic skyline, the
Magnificent Mile, and numerous cultural attractions. The residential environment of Central
Chicago consists of luxurious vertical structures in the Near North and a recent wave of
redevelopment in the old industrial infrastructure in the Near South. The Loop is the commercial
center of Chicago and is mostly non-residential.

36

Extending from Central Chicago are three prominent regions on the semi-periphery of
the Loop: the North Side, the South Side, and the West Side. The North Side has the highest
concentration of residents in the city, and contains numerous public parks and beaches stretching
along Lake Michigan. This area underwent a process of gentrification beginning in the early
1990’s as many of the Eastern European ethnic enclaves were replaced with an upper middle
class (Reiff, et al, 2005). The North Side’s features include the home of the Chicago Cubs, the
Lincoln Park Zoo, and Boystown. The community areas comprising the North Side have some of
the wealthiest neighborhoods in the country and have the highest average income in Chicago
(Keating, 1998).
The South Side is a more industrial, poorer, and ethnically diverse area of the city. The
South Side has a higher ratio of single-family homes and contains many of the city's original
industrial spaces (Randall, 1999). The South Side has two of Chicago's largest public parks.
Jackson Park, which hosted the World's Colombian Exposition in 1893, is currently the site of
the Museum of Science and Industry. The park stretches along the lake front, linking the
neighborhoods of Hyde Park and South Shore. Washington Park is connected to Jackson Park by
the Midway Plaisance and is located just west of the University of Chicago. The South Side has
an ethnically diverse population with many black and Eastern European neighborhoods (Reiff, et
al., 2005).
The West Side, consisting of community such as Austin, Lawndale, Garfield Park, West
Town, and Humboldt Park, has two distinct sections with very different socioeconomic
landscapes. The community areas to the far west, particularly Garfield Park and Lawndale, have
had long-term socioeconomic problems, a history of crime and segregation, and infrastructural
concerns (Randall, 1999). Black and Hispanic populations mainly occupy the spaces of Garfield

37

Park and Lawndale. The West Side neighborhoods closer to downtown are in the process of
undergoing significant redevelopment as many of the old ethnic enclaves- Ukrainian Village,
Greek Town, and Wicker Park for instance- are being remodeled and rebuilt to house the modern
urban upper middle class (Channick, 2008).

Figure 1. Map of Chicago community area regions. This map depicts the nine regions the
community areas of Chicago fall within. The community area region is the most general
boundary units for the city. Map design by Nicholas A. Perdue, 2012. Data provided by Chicago
Metro Agency of Planning and by the City of Chicago GIS department. For interpretation of the
references to color in this and all other figures, the reader is referred to the electronic version of
this thesis
Extending further from the Central Region are two regions with a mixture of urban and
suburban characteristics- the Northwest Side and the Southwest Side. The city's Northwest Side
38

is a residential section populated primarily with people of Eastern European decent (Randall,
1999). This region of the city has many Ukrainian, German, and especially Polish
neighborhoods that have not experienced the same rates of redevelopment or demographic
change as other older immigrant neighborhoods in the city (Channick, 2008). The Southwest
Side, like its neighbor to the east, is an industrial, lower class, and ethnically diverse region in
the city. The Southwest Side includes the community areas of Garfield Ridge, Clearing, Archer
Heights, McKinley Park, New City, Chicago Lawn, and Gage Park (Reiff, et al., 2005). These
areas have a mixture of relatively segregated Hispanic, Eastern European, and black
neighborhoods (Channick, 2008). Many of the Irish, Polish, and Lithuanian neighborhoods in the
area are experiencing demographic changes as recent immigrants, primarily people of Puerto
Rican and Middle Eastern descent, are moving in as older generations move towards the western
suburbs (Keating, 1998). The Southwest side also includes some of the more dangerous
neighborhoods in the city, particularly the community area of Englewood, which has among the
highest crime rates in the country (Reiff, et al., 2005).
The remaining three regions lie on the geographic outskirts of Chicago and are
characterized by either suburban residential or industrial sectors (Reiff, et al., 2005).The Far
North of Chicago is a region defined by O’Hare Airport but also includes the areas of Rogers
Park, Lincoln Square, and Edison Park. The Far North includes areas of old and new immigrant
communities, Asian enclaves, high-end residential neighborhoods, LGBT communities, and
recently gentrified mixed-use urban spaces (Reiff, et al., 2005). The Far Southwest Side is a
mostly residential section of the city comprised of mainly Irish and black neighborhoods (Reiff,
et al., 2005). This region of the city features such areas as Beverly, Ashburn, and Washington
Heights and contains some of the most ethnically diverse community areas (Channick, 2008).

39

The Far Southeast Side is a mostly industrial section of the city mixed with some lower income
residential neighborhoods (Reiff, et al., 2005). The historical Pullman District, where many early
labor and social activist battles were fought, including the formation of the first labor unions,
highlight the Far Southeast of Chicago (Pierce, 2007). The Far Southeast is the poorest and most
racially segregated area of Chicago with primarily black residents. High amounts of violence,
gang activity, fragmented neighborhoods, and decaying infrastructure characterize this part or
Chicago (Reiff, et al., 2005).

3.5 Modern Urban Structure of Chicago
The boundaries and borders of spaces have a direct effect on the residents; perhaps not in
a determining cause for actions and experiences, but in an indirect way of determining
differential power and social ties. A discrepancy between the administrative boundaries and the
residentially defined social boundaries clearly exist, and thus, is a cause of distortion and bias in
the study of population distributions within the neighborhood units (Bruhn, 2009).
Administratively, neighborhoods are seen as territorial-bounded entities (Bruhn, 2009). In
reality, people live and experience urban life in a series of overlapping social networks.
Geospatial analysis which concentrates on the bounded units simplifies and weakens the lived
experience of the residents. In Chicago, the community areas have established names, static
borders, and a built in conception of residential patterns.
The form, character, and essence of a neighborhood are subject to how the neighborhood
boundaries are defined. Community area boundaries and borders influence, to some extent, the
lived experience of the residents. To a greater extent, the boundary subdivisions in Chicago
define the study and classifications of space based on socioeconomic and demographic

40

distributions. The boundaries of the Chicago community areas rely on units of aggregation with
links to the old political and ethnic structures established during urban growth of the 19th and 20th
century, on features of the built environment such as roads and parks, and on the natural
boundaries of the Chicago River and Lake Michigan. Such aggregate proxies may be
inappropriate for learning about the micro-processes of the urban space and for realistic
visualization of the population distribution (Bruhn, 2009).

Any city-dweller knows that most neighborhoods don't have stark boundaries. Yet on
maps, neighborhoods are almost always drawn as perfectly bounded areas, miniature
territorial states of ethnicity or class. And maps showing perfectly homogeneous
neighborhoods are still published today, in both popular and academic contexts alike
(Rankin, 2009 pp 2).

3.6 Population Density in Chicago
In Chicago, the representations of conventional metrics lead a map reader to believe that
the most crowded areas are those in the initial zone around The Loop- in the community areas in
the Near North and the Near South- and the amount of space steadily increases moving away
from Lake Michigan towards the suburban neighborhoods. This conception is the direct result of
modeling populations using planar surfaces, boundary enumerations, and the utilization of
choropleth maps. This impression relies on boundary units and ignores the vertical extent of the
residential environment distributed across the urban spaces of Chicago. The conventional
measurement of population density fails to capture the phenomena of urban living in the 21 st
century. To understand the spatial distribution of people in contemporary Chicago, the
movement away from population density is essential.

41

Figure 2. Choropleth Map of Population Density in Chicago. This map depicts the population
per square mile of ground space within the community area boundary. The Choropleth technique
assigns a singular value to all spaces within each community area that change abruptly at the
border. This map uses a sequential color scheme ranging from light for low values to dark for
high values. Map design by Nicholas A. Perdue, 2012. Data provided by United States Census
Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS department;
background image provided by Google Earth.
.

42

3.7 Chicago Summary
The development of Chicago was one of initial shock- both in terms of population and
land acquisition. The urban space of Chicago transformed after the Great Chicago Fire in 1871
and gave the city a chance to restart its developmental process and modernize the unorganized
and quickly growing space into a city of advanced transportation, organization, and architecture.
The post-fire redevelopment of the downtown urban core and the competition with New York
towards modernism created an organized space with planned parks, neighborhoods, and vertical
buildings.
The Chicago School, using a lens of critical urban theory and the ideas of urban ecology,
developed a boundary system in 1925 to define and separate the city with a static system of
boundaries and borders. The community area boundary system has remained consistent since
formation and continues to dominate the geographic perspective of the city. Population
distributions in Chicago are often represented as homogenously distributed collections of people
within the community area enumerations. This construction of the living space in Chicago,
rooted in the historical geographic context and development of the city space, fails to account for
the modern urban space of vertical residential buildings and spatial distribution of people. The
backdrop of Chicago provides a fitting study area for the exploration of modern boundary
systems, the deconstruction of conventional population density, and the modeling of vertical
spaces in population analysis.

43

Chapter 4
METHODS

4.0 Data
The data used for this research are publically available and were acquired between
October of 2010 and March of 2012. The 2010 Census population data were obtained from
American Fact Finder and aggregated at the census block geographic level. The data are from the
summary file 1 (SF1) data distribution. Building footprints, boundary units, parks, streets, and
other natural features are found on the City of Chicago Geographical Information website, the
Cook County Geographic Information website, and the Illinois Natural Resource Geospatial Data
Clearinghouses- all downloaded in the format of ESRI shapefiles. The building vertical extent
and number of floors is an attribute in the building footprint dataset. Building heights were
confirmed using LIDAR as part of the Illinois Height Modernization Initiative in 2008. All
basemaps and aerial imagery are from ESRI’s ArcServer web application and Google Earth.
Zoning codes for the city are available from the Chicago Metropolitan Agency of Planning data
deposit. The zoning codes are also from 2008 and joined to the building footprints to determine
building usage. Historical Maps of Chicago are found at the University of Chicago Map Library.
All photographs used in descriptive analysis were taken by Nicholas A. Perdue between May and
October, 2011 unless otherwise noted.
This research uses population and demographic data from summary file 1 (SF1) that
includes income and racial distributions at the census block level. The methods in this chapter
use population data from the census block level of measurement. The census block measurement
allows for the smallest spatial units in the source data. There are over 16,865 census blocks in

44

Chicago with a population of 2.7 million people, distributed in the living spaces of more 400,000
residential structures throughout the city.

4.1 Overview and Goals
This chapter describes new methods to characterize and visualize urban settlement
patterns in new ways. Conventional population density approaches fail to capture the complex
patterns of people living in contemporary cities. The methods in this chapter provide alternative
techniques to represent the complicated patterns of human concentrations and vertical
environments in contemporary cities. Conventional analytical techniques rely on the established
boundary units and represent Chicago’s population in a simplified and typically antiquated
manner. Experimentation in the methodological approach allows for the creation of maps and
visualizations designed to better characterize the phenomenon of people in urban space. The
methodologies for this section fall into two broad categories, geoprocessing methods and
cartographic methods. The two approaches function together to create and classify a new metric
of livable space in vertical cities and to visualize human concentrations in a unique way. The
methods are organized and discussed in the following sections.

1. Section 4.2 describes the creation of a new ‘personal space’ metric that evaluates the
living spaces of Chicago.
2. Section 4.3 introduces a new classification scheme designed to categorize the different
types of living spaces in Chicago.
3. Section 4.4 explains a technique to understand how demographic profiles change based
on how living space is defined.

45

4. Section 4.5 discusses the development of an original visualization approach designed to
reveal the population structure of Chicago.

4.2 Personal Space Model
The vertical residential environment of the contemporary creates a spatial extent that is
not characterized by conventional population density. Conventional population density only
analyzes the surface of the earth, confusing places of high density and places of crowded living
spaces. The livable spaces in cities are defined as places within the residential structure where
people actually live. The numeric value the personal space metric aims to calculate is the amount
of space each person has within the built housing environment of the city. The goal of this
method is to create a metric which shifts the concept of crowdedness from conventional
population density into the measurement of personal space.
The first step in the personal space model is a binary dasymetric approach separating
residential and non-residential structures throughout the city. This method eliminates all the nonresidential structures based on urban zoning codes attached to the building. The binary approach
deems all buildings zoned for residential as suitable for living and all other buildings as
unsuitable, masking out the spaces where people do not live (Maantay et al., 2007). Mix use
buildings- structures with both commercial and residential zoning designations- complicate the
binary analysis. Mix use buildings are common in all cities, particularly in redeveloped urban
space, areas along major artillery roads, and places near natural features such as parks and
waterfronts (Grant, 2002).

46

Figure 3. Personal Space Flowchart. The personal space model uses three input variablesbuilding footprints, census block populations, and zoning codes to create a measure of the space
each person in the city has within residential buildings. A binary and areal interpolation
dasymetric method is utilized to allocate populations to only buildings. Model design by
Nicholas A Perdue.

The binary approach counteracts the limitations of bounded areal units by assigning only
specific buildings as suitable for living. In the binary step, mix use buildings, are considered
suitable for inhabitance. The binary approach assumes a homogenous residential space in the
buildings and assigns a singular value to the space, not accounting for variation within buildings.
This variable of residential suitability is represented as a 0 or 1 and is named residential
buildings.

47

The second step in the personal space method is to calculate the total amount of the
livable space for each of the residential buildings. The amount of livable space is calculated
using the product of the square footage of the building footprint, the number of stories for that
building, and the binary value of the residential buildings variable. This step calculates the total
living space for each building in the city, represented with a variable livable space. The total
living space for mix use buildings is calculated by assigning 40% of the livable space to non
residential and 60% to residential. This allocation follows models used in urban development
simulation experiments (Waddell, 2003). The final step is an areal interpolation approach
allowing for the transformation of the source data set into a target data set (Mennis 2003). The
source data are the raw census population totals aggregated at the census blocks; the target data
are livable space. The areal interpolation method aggregates the population totals of the census
block to what percentage of the cumulative livable space each building possesses. For example,
if one building accounts for 25% of the cumulative livable space of a census block, 25% of the
population is allocated to that specific building. Areal weighting is a simple interpolation method
which allocates the summarized population data according to the proportional area of the zones
in the target data (Langford, 2003).
The three steps create the personal space metric, the amount of livable space each person
in the city has in residential structures. The personal space metric assigns a value to every
residential structure in the city with the amount of space per person.
The personal space metric is designed not to be a measure exclusive to Chicago, but
rather a calculation that can be performed in any city. The metric requires three inputs- zoning
codes, building footprints, and census data; most major cities with extensive vertical spaces
should have this data. The metric is aimed at being easily repeatable regardless of geography.

48

Figure 4. Personal Space Equation. The personal space model is designed to calculate the total
residential living space within a census block and allocate the raw population back into buildings
based on how much space is available within the vertical residential environment.

4.3 Classification of Personal Space
The goal of the personal space approach is to create a new metric to understand where
conventional measurement of population density fails to accurately characterize the
contemporary urban space. The classification of this metric uses a bivariate legend design to
illustrate the relationship between conventional population density and personal space,
highlighting the areas of Chicago that are misrepresented by the conventional measurement. The
bivariate legend divides the range of conventional densities at the census block level into three
quantiles plotted on the x-axis, creating an index of high, medium, and low conventional density
values. The personal space metric is plotted on the y-axis and also divided into three quantiles,
creating an index of high personal space, medium personal space, and low personal space
classifications.
Combining personal space and conventional density on a three by three grid and
symbolizing with a diverging color scheme shows the relationship of conventional density and
personal space. A diverging color scheme enables the interactions between the two
measurements to be highlighted in the four corners of the legend. The classification scheme is
applicable at every level of measurement, from the individual building to the community area. At
spatial units greater than the individual buildings, the bivariate design uses the median value of
personal space within the unit borders to position that space within a quantile for classification.

49

Figure 5. Bivariate Legend-The bivariate legend is designed to show the traditional population
density on the y-axis and the new personal space model on the x-axis. This will allow for the
spaces where population density is accurately or inaccurately characterizing the spaciousness or
crowdedness of the residential environment. The legend uses a diverging color scheme on both
axes to emphasize the corners. Legend design by Nicholas A Perdue, 2011.
In the bivariate legend, the four corners indicate spaces where the high and low classes
interact. The lower left corner of the legend, represented with pink, indicates the areas with a low
conventional population density and a low amount of personal space. The upper right corner of
the legend, represented with green, shows the areas with a high conventional population density
and a high amount of personal space. The upper left classification, represented with purple,
shows the spaces of high density and low personal space. The lower right classification
represented with orange, illustrates the spaces low in density and high in personal space. For the
remainder of this thesis, these four classes will be referred to by the assigned color.

50

4.4 Demographic Profiles of Chicago
A function of the personal space measurement is the ability to evaluate the residential
spaces of Chicago in a new and informative way. Utilizing personal space in analysis allows for
a better understanding of how people are spatially distributed across different urban
environments. In order to better understand this distinction and to explore the implications of the
new metric, this section examines how demographic distributions change when the definition of
the living space is changed.
For this method, racial or ethnic self-identification is divided into five categories for
analysis and representation: white, black, Hispanic, Asian, and other.
Exploring demographic profiles illustrates how the new metric reveals a more in-depth
picture of urban life. The comparison of conventional and new metrics through the lens of racial
breakdowns of different spaces in Chicago shows changes with the enhanced methodologies.
First, all conventional high density spaces are contrasted the green and the purple classifications.
The green and purple classifications represent the high and low measure of personal space within
the high classification of conventional density. Similarly, all conventional low density spaces
are contrasted with orange and the pink classifications. The orange and pink classifications
represent the high and low measure of personal spaces within the low classification of
conventional density. The comparison of the different spaces shows both percent of total
population for the different classifications as well as raw population numbers. The demographic
profiles are collected from the census block measurement.

51

4.5 Individual Representation Method
The goal of the individual representation method is to create a series of maps that
represent as precisely as possible the residential patterns in the urban space. This method is
rooted in conventional dot density approaches but with a focus on micro-scale representation of
the individual people in cities. In this method, dots are placed in the residential building only to
maximize precision in location. Each dot represents one person to increase the accuracy in
representation and to avoid the generalizations stemming from representing many people with a
singular visual. The boundary demarcations of census blocks and neighborhoods are removed in
the final maps to show how individual people occupy the built environment of the city.
In the individual representation method, the goal is to place the visual variable as close to
the location of the entity being mapped, to the building where the individual lives. A Poisson
distribution determines the population of each building based on the expected population of each
building. The expected population for each building is derived from the personal space metric.
The personal space metric creates a measurement that quantifies the amount of residential space
in each building per person, the inverse being the amount of expected people per building. From
the expected intensity, the Poisson distribution allocates individuals into the building to be
represented in a dot density map. A Poisson distribution allows for a level of variability within
the buildings and assures that buildings are not homogenous units across the census blocks,
ensuring that every building in the census block is not assigned individuals based on the common
value of average personal space.
The dot density method for this series of maps represents each individual in the city with
a singular dot confined to a residential building. The dots use a symbology in which the hue of
the dot represents one of the five census race categories- white, black, Hispanic, Asian, or other.

52

Chorodots allow for the representation of demographic attributes while maintaining similar size
and spatial locations across all observations (MacEachren, 1990). The chorodot technique allows
for the representation of different groups of individual people filling both the planar and the
vertical spaces of the urban environment with a single visualization.

4.6 Summary
This section illustrates a variety of techniques aimed at providing a more detailed
representation of the phenomenon of contemporary urban residential life. The approach outlined
in this section combines geoprocessing methods and cartographic methods to create, classify, and
represent personal space and populations in the built environment of the contemporary city. The
vertical space is a feature of urban residential structures that is not accounted for with the
conventional measurements and visualizations. The personal space index and the bivariate
classification are methods to model the vertical space of cities as an important part of the
residential environment. A fine scale cartographic approach of individual representation allocates
the demographic characteristics of individuals to be visualized at the building level and explores
the microstructure of the built environment in Chicago. The methods presented in this section
work together to achieve a singular goal; to develop a more developed understanding how
humans are concentrated in contemporary urban spaces with a series of new metrics for analysis
and visualization.

53

Chapter 5
RESULTS

5.1 Overview and Goals
This chapter describes the results of the methods introduced in the previous chapter. The
methods presented deconstruct conventional approaches used in urban population analytics and
create new ways to characterize and visualize urban settlement patterns. Reviewing the statistical
and cartographic outputs from each of the methods illustrates the ways in which conventional
methodologies fail to capture the complex pattern of how people live in contemporary cities. The
results are organized and discussed in the following sections.

1. Section 5.2 describes how the ‘personal space’ metric provides an alternative to
population density.
2. Section 5.3 discusses how the bivariate classification scheme illustrates the confusion
between of density and crowdedness.
3. Section 5.4 examines how demographic profiles of census blocks dramatically change
under the new metric of personal space.
4. Section 5.6 judges and provides examples of the highly focused dot density technique in
the individual representation method.

5.2 Personal Space Model
The personal space model derives from the vertical extent of the built environment in
contemporary cities that is not characterized by conventional population density. A map showing

54

the number of high rise buildings in Chicago in 2008 demonstrates why modeling the vertical
space in contemporary cities is a valuable.

Figure 6. Map of buildings over seven stories in Chicago, 2008. The number of high rise
buildings in modern Chicago. There has been a significant increase in the number of high rise
buildings since 1933 when the University of Chicago map was made depicting the vertical
spaces in the city. The distribution of buildings remains similar, however, with the majority of
high rises in the Loop and to the north and south along Lake Michigan. Map design by Nicholas
A. Perdue, 2012. Data provided by Chicago Metro Agency of Planning and by the City of
Chicago GIS department.
55

The personal space metric calculates the amount of living space each person has within
the residential structures of across the extent of Chicago. Personal space is a measurement of
square footage of residential floor space per person. The metric assigns a value of space per
person to every residential structure in Chicago. There are two main limitations to the personal
space method.
In the personal space metric, the first step is to calculate a total amount of residential
space per building within each census block. This calculation includes all common areas of
buildings including hallways, communal laundry facilities, storage rooms, and entrance ways.
During the areal interpolation method, the census totals are allocated to a total that includes this
extra not necessarily residential space. A fraction of the personal space measurement, depending
on the type of structure, is common space within the building and not explicitly personal
residential space. Common space within buildings is a part of the personal space metric, common
space outside the residential structure such as courtyards and parking lots is not. This limitation
exposes a taxonomy of spaces in cities. In the urban environment, there are private, semi-private,
semi-public, and public spaces. Private and semi-private can be described as places within the
residential structures; public and semi-public as places outside. The private and semi-private are
both included in the personal space metric.
The second limitation is a product of the unit of source data, the census block population
aggregations. Each individual within a given census block is allotted a percentage of the total
residential floor space from within that specific census unit. The areal interpolation relies on the
assumption that population is distributed uniformly within the target zones (Maantay, 2007). The
buildings are a percentage of the total livable space within the population unit; individual
building variation within the block cannot be achieved with this approach. As a result, all the
56

buildings within a census block will have an identical measurement characterizing the amount of
space per person. Both of these limitations add a degree of uncertainly to the model.
Regardless of the limitations, the personal space method is effective metric of human
settlement patterns in cities. The personal space metric eliminates all the places in the city that
are public, commercial, industrial, natural, or infrastructural spaces and classifies these areas as
non-residential. The dasymetric techniques to allocate people to a three dimensional environment
successfully combats the conceptual limitations of conventional urban population density
mapping. The personal space measurement shows not how many people live in an area, but how
people live within that area. The personal space metric more precisely answers the question of
how populations are distributed across the built environment of the contemporary urban space.

5.3 Classification of Personal Space
The bivariate choropleth design allows for the spaces in Chicago where complex
relationships of conventional density and personal space exist to be clearly illustrated. The
classification method divides the conventional density and the personal space values into
quantiles creating classes of high, medium, and low values for each variable. The quantile
classification allows for the range of the values and the size of the city to determine values of
high, medium, and low rather than having concrete break points. The quantile classification
allows for the method to be transferred to other cities where the perception of high and low
amounts of space could be different. The ranges of population densities and personal spaces will
certainly be different between different cities.

57

Figure 7. Bivariate Legend. The bivariate legend creates four unique classifications of space at
the four corners of the (x, y) grid. These four classifications, referred to by color, are the spaces
of interest in Chicago analyzed in this section. Legend design by Nicholas A Perdue, 2011.

Perceptions of space in Chicago will be different than that in New York, San Francisco,
or Denver. Utilizing quantiles allows for the break points to be determined by the data of the city
itself, not by the data of another city. In other words, high personal space in Chicago is classified
as high if it is in the upper third of the range of personal spaces available in Chicago. Assigning
values of high, medium, and low to both variables is done so relative to other spaces within the
city of study. The classification of census blocks rather than individual people illustrates the
differences between the high density and low density classification more effectively. Ideally,
census blocks should be of similar populations. Showing the differences in populations between
blocks gives a sense of the variations in the usage of the city space.

58

The diverging color scheme illustrates four areas of interest located in the four corners of
the legend grid- the green, purple, orange, and pink classifications. The four corners of the
legend indicate the space where the classifications of high and low for the two variables interact.
The first set of results use the bivariate design to classify the community area level of
measurement to understand a general spatial pattern of Chicago at an identifiable and commonly
recognized geographical extent. The average personal space is calculated for each of the seventyseven community areas in Chicago by taking the median value of all buildings contained within
the boundary. The community area are plotted and classified on the bivariate grid creating a map
of personal space juxtaposed with conventional population density.
The pink and the green community areas indicate spaces where the conventional
population density and the new metric of personal space most deviate from each other. The green
areas, for example, are in the upper third in conventional population density as well as in the
upper third in personal space. A space high in convention density indicates a degree of
crowdedness only if the surface space of the community area urban structure is considered; the
high amount of personal space shows the more realistic picture of these parts of the city- a
relatively spacious residential environments. Conversely, the conventional density measurement
for the pink class indicates a degree of spaciousness but the personal space metric shows low
amounts of space within the residential structure.

59

Figure 8. Bivariate Map of Chicago. The interaction of personal space and conventional
population density is mapped across the extent of Chicago, at the community area level of
measurement, using the bivariate legend design. The four classes in the corners of the legend
represent unique urban residential spaces that are not characterized properly with the
measurement of population density. Map design by Nicholas A. Perdue, 2012. Data provided by
United States Census Bureau, Chicago Metro Agency of Planning and by the City of Chicago
GIS department; background image provided by Google Earth.

The green classification represents ten of the seventy-seven community areas in Chicago.
The community areas are listed by name, conventional population density in square feet per

60

person, and the personal space in square feet in table 1. The ten green community areas are
defined by a high amount of personal space in residential structures in community areas where a
high number of people reside per unit of ground area. The community area of Lakeview, located
along Lake Michigan in the North Side of Chicago, shows an average value of 1,133ft 2 of
personal space while the allotted amount of ground area per person is only 812ft 2 per person.
This indicates a high vertical extent to the community area and a high concentration of
multistory residential structures.

Table 1. Green community areas. This table shows the amount of space each person within the
green classified community areas have within residential buildings and within the planar surface
of the community area. Lakeview shows that there is more residential space area than total space
within the community area boundary.

The green classification shows a spatial distribution primarily along Lake Michigan both
north and south of the Loop. A majority of the high-rise residential structures are located along
the lakefront in the community areas classified green. There is a strong visual correlation
61

between the green community areas in Figure 10 and the distribution of high rise residential
buildings in Figure 8. The green class illustrates how a conventional population density
measurement is a poor representation of the space people live in within the vertical urban setting.
The pink class represents eight of the seventy-seven community areas in Chicago. The
community areas are listed by name, conventional population density in square feet per person,
and personal space in square feet in table 2. A low amount of personal space within residential
structures in sparsely populated community areas characterizes the pink community areas. The
community area of South Deering, located in the Far Southeast of Chicago, has an average of
358ft2 of personal space while the allotted amount of ground area is 17,330ft 2 per person. This
interaction indicates a collection of smaller, crowded residential dwellings in spaces where high
amounts of the community area are filled with non-residential buildings and land uses.

Table 2. Pink community areas. This table shows the amount of space each person within the
pink classified community areas have within residential buildings and within the planar surface
of the community area. South Deering has a large amount of space in the community area
leading to a low density measurement. The residential spaces however are quite crowded.

62

Ten of the seventy seven community areas in Chicago fall into the orange classification.
These community areas are listed by name, conventional population density in square feet per
person, and personal space in square feet in the table 3. The orange community areas are defined
by a high amount of personal space in residential structures in conventionally low density
community areas. In the case of the orange measurement, the conceptual ideas of population
density as a measure of crowdedness is preserved- a low population density equates to a spacious
residential environment.

Table 3. Orange community areas. This table shows the amount of space each person within
the orange classified community areas have within residential buildings and within the planar
surface of the community area. Forest Glen is spacious both inside and outside of the residential
environment.

The community area of Forest Glen, located in the Far North of Chicago, shows a
suburbanization land use pattern with an average value of 760ft 2 of personal space in a
community area with 4,462ft 2 of ground area per person. This pattern of residential development

63

can be characterized as suburban where single family houses are positioned on relatively large
parcels of land. The spaces both within the home and within the community areas in the orange
classification depict spacious environments.
The orange classification shows a spatial distribution of community areas primarily
located on the perimeter of Chicago in the spaces that were annexed towards the end of the
growth and development period in the early 20th century. The exception is the Near South Side
community area which experiences a different land use and recent development pattern.

Table 4. Purple community areas. This table shows the amount of space each person within
the purple classified community areas have within residential buildings and within the planar
surface of the community area. South Lawndale is crowded both inside and outside the
residential buildings.

The purple classification represents nine of the seventy-seven community areas in
Chicago. These community areas are listed by name, conventional population density in square
feet per person, and personal space in square feet in the table 4. The purple community areas are
defined by a low amount of personal space in residential structures in conventionally high

64

density community areas. In the purple class the conceptual ideas of population density as a
measure of crowdedness is preserved.
The community area of South Lawndale, located in the Southwest Side of Chicago,
stands out in this grouping, with an average value of 239ft2 of personal space with the allotted
amount of ground area per person at 1,368ft 2. This pattern of residential development is
characterized by tightly built urban single family houses and small apartment complexes
positioned on relatively small parcels of land. The purple class, both within the home and within
the community areas, is portraying a crowded space.
The spatial distribution of the purple classification shows a pattern of community areas
primarily located on the western edge of Chicago in the spaces that were annexed fairly early in
the suburbanization process. The purple community areas are the spaces where many of the early
immigrant neighborhoods formed in the years of post-fire redevelopment and growth.
The bivariate classification scheme effectively shows the juxtaposition of personal space
and conventional population density. Analytics at the community area geography allows for
statistical comparisons, the emergence of general patterns, and descriptive at a recognizable
scale. The community areas, however, represent a system of boundaries that is outdated and
generalizes too much of the urban space into homogenous bounded units. The census block is a
more appropriate scale for in depth analysis on the relationship between conventional density
and space. At the census block level, the relationships between demographics profiles at
conventional and new metrics of urban space can be explored.

65

Figure 9. Bivariate map of Chicago at the census block geography. The Census block
geography is the units of measurement used for the demographic analysis illustrating the
differences of high and low density spaces when residential spaces inside buildings are
accounted for. The map is overlaid on the Google Earth interface. Map design by Nicholas A.
Perdue, 2012. Data provided by United States Census Bureau, Chicago Metro Agency of
Planning and by the City of Chicago GIS department; background image provided by Google
Earth.

66

5.4 Demographic Profiles of Chicago
The purpose of this section is to use the personal space measurement to explore how
demographic distributions change when the definition of the space is altered. The racial classes
in this section are divided into five categories: white, black, Hispanic, Asian, and other.
In total, there are 5,592 census blocks with the conventional high density measurement in
Chicago with the following demographic profile: white 31%, black 24%, Hispanic 28%, Asian
4%, and other 13%. Of Chicago’s 2.7 million people, 1.6 million (60%), live in the conventional
high density blocks spatially accounting for one-third of the space in the city. With the improved
measurement of personal space, the demographic profiles of the conventional high density space
of Chicago change dramatically.
Investigation of the green classification reveals a clear shift in the demographic landscape
of the census blocks. The white race makes up 31% of the total population of people in
conventional high density areas, roughly 500,000 people. In the green class, white accounts for
57% of the population. The green classification makes up 1,238 census blocks, just over 7% of
the city space.

67

Figure 10. Demographic profile of conventional high density census blocks. The
demographic profile for the five races in the study from within all Census blocks classified as
high density. The high density blocks consist of those in the upper third in the number of people
per square foot of ground space in the block. Map design by Nicholas A. Perdue, 2012. Data
provided by United States Census Bureau, Chicago Metro Agency of Planning and by the City of
Chicago GIS department.

68

Figure 11. Demographic profile of green census blocks. The demographic profile for the five
races in the study from within all Census blocks classified as high density, high personal space.
The green classification indicate the Census blocks in Chicago in the upper third in the number
of people per square foot of ground space in the block and in the upper third of square footage of
space per person inside buildings. Map design by Nicholas A. Perdue, 2012. Data provided by
United States Census Bureau, Chicago Metro Agency of Planning and by the City of Chicago
GIS department.
The total population for those inhabiting the green blocks is 400,000 with 225,000 self
identifying as white. The green class represents the vertical residential spaces of Chicago, places
with spacious living environments, and arguably the wealthiest and most exclusive space in the
city. The green census blocks in Chicago are dominated by the white race.
The jump in the percent white in the green classification is at the expense of the Hispanic
and other demographic classes. As the percent white rises from 31% to 57%, the percent
69

Hispanic drops from 28% to 9% and the percent other drops from 13% to 3%. Of the 450,000
Hispanics that live in the conventional high density blocks of Chicago, only 35,000 live in the
green class. Of the approximate 210,000 high density residents from the other classification –
including American Indians, Alaska Natives, Native Hawaiians, Pacific Islanders, ArabAmericans, Persian-Americans, and Indian-Americans- only 12,000 live in the green. The
percents Black and Asian stay approximately the same between the two classifications.
The improved personal space metric exposes a different set of patterns in the purple
classification when compared to conventional high density spaces. The purple class has a total of
2,286 census blocks with a population of 800,000 people. Spatially, the purple blocks account for
17% of the space in Chicago. In terms of population, the purple accounts for 30% of the people
in the city. Of the 800,000 people in the purple classification, 24%, or 190,000, self-identified as
white in the 2010 census. Comparing the green and purple class, there are 30,000 more people of
the white race that live in the green 7% of the city space than there are in the purple 17% of
Chicago. The percent white from the conventional high density class to the purple class drops
from 31% to 24%.
The decrease in white population in the purple class relates directly with the increase in
the percent Hispanic from 28% to 37% and the increase in percent other from 13% to 19%. Of
the 800,000 people living in the purple blocks, 300,000 are Hispanic. Of the 450,000 Hispanic in
conventionally high density spaces of Chicago, 300,000 are in the purple class whereas only
35,000 are in the green locations. The other racial category shows a similar pattern; of the
210,000 people in the conventionally high density blocks, 150,000 live in the purple spaces
while only 12,000 live in the green.

70

Figure 12. Demographic profile of purple census blocks. The demographic profile for the five
races in the study from within all Census blocks classified as high density, low personal space.
The purple classification indicate the Census blocks in Chicago in the upper third in the number
of people per square foot of ground space in the block and in the lower third of square footage of
space per person inside buildings. Map design by Nicholas A. Perdue, 2012. Data provided by
United States Census Bureau, Chicago Metro Agency of Planning and by the City of Chicago
GIS department.

The conventional low density blocks in Chicago have a very different residential
demographic structure than the conventional high density blocks. To illustrate the demographic
landscape of these spaces, the comparison between the orange and the pink classifications are
explored at the census block level.

71

Figure 13. Demographic profile of conventional low density census blocks. The
demographic profile for the five races in the study from within all Census blocks classified as
low density. The low density blocks consist of those in the lower third in the number of people
per square foot of ground space in the block. Map design by Nicholas A. Perdue, 2012. Data
provided by United States Census Bureau, Chicago Metro Agency of Planning and by the City of
Chicago GIS department.

There are 5,606 conventionally low density blocks in Chicago with the following
demographic profile: white 45%, black 40%, Hispanic 9%, Asian 2%, and other 4%.
Conventional low density blocks contain a total of 400,000 people, 15% of the total population
of Chicago. Whereas the conventional high density blocks were mostly spatially distributed
along the Lake Michigan coast and in the interior of western Chicago, the conventional low
density blocks are mainly concentrated in the Far Northwest and the Far Southwest. The

72

improved measurement of personal space noticeably changes the demographic profiles of the
spaces conventional measurements deem as low density. The shift in the demographic geography
is quite apparent in the orange blocks. The white population makes up 45% of the total
population of the conventional low density areas, 180,000 people. In the orange classification,
the percent white jumps to 53%.

Figure 14. Demographic profile of orange census blocks. The demographic profile for the
five races in the study from within all Census blocks classified as low density, high personal
space. The orange classification indicate the Census blocks in Chicago in the lower third in the
number of people per square foot of ground space in the block and in the upper third of square
footage of space per person inside buildings. Map design by Nicholas A. Perdue, 2012. Data
provided by United States Census Bureau, Chicago Metro Agency of Planning and by the City of
Chicago GIS department.

73

The orange classification makes up 17% of the city space with a total of 2,853 census
blocks. This orange space, as a whole, contains just over 200,000 people, about 7% of the
population of Chicago. Of the 180,000 of the white race living in all the conventionally low
density areas of Chicago, 105,000 live in the orange class. The orange classification represents
the primarily suburban single family homes- places with spacious living environments on
relatively large parcels of land. The residential environment is spacious both in and outside the
home. The orange census blocks are largely dominated by the white race.
The high personal space classification, regardless of the conventional density
measurement, consists of 5,650 census blocks containing a total of 800,000 people. The white
race makes up 55% of the total population in this space, 440,000 individuals. Chicago, as a city,
has a population that is 32% white, a total of roughly 865,000 individuals. More than half of
those identifying as white in Chicago live in the most spacious residential environments of the
city. Conversely, only 56,000 of Chicago’s 785,000 Hispanic residents, just over 7%, live in the
most spacious parts of the city. There is a clear distinction in the demographic landscape of
Chicago and it is based not on how many people live within the boundary unit, but on how
people live within the built environment and residential structures of a space, in the degree of
crowdedness or spaciousness inside the home.
There are 1,211 census blocks that fall into the pink classification. The pink space in
Chicago, though equaling 7% of the land area accounts for less only 3% of the total population
with just over 82,000 individuals. The percent black and percent Hispanic rises in the pink
spaces as the percent white decreases. The pink areas can be characterized as mostly fragmented
spaces along the industrial and transportation infrastructure of the city; places where few live but
those that do reside in small and crowded residential structures.

74

Figure 15. Demographic profile of pink census blocks. The demographic profile for the five
races in the study from within all Census blocks classified as low density, low personal space.
The pink classification indicate the Census blocks in Chicago in the lower third in the number of
people per square foot of ground space in the block and in the lower third of square footage of
space per person inside buildings. Map design by Nicholas A. Perdue, 2012. Data provided by
United States Census Bureau, Chicago Metro Agency of Planning and by the City of Chicago
GIS department.
Of the 2.7 million people living in Chicago, 40% are living in buildings with a low
amount of personal space. The percent Hispanic for all low personal space area, regardless of the
density of the surrounding block, is 32%-, a total of 355,000 individuals. Six Hispanic people
live in a low personal space building for every one that lives in a high personal space building. A
total of 290,000 white individuals live in low personal space areas across the city, or two for
75

every three living in a spacious residential environment. Of the 891,000 Black residents of
Chicago, 225,000 live in the spacious residential environment while 300,000 live in crowded
residential spaces. An individual classified as other in the demographic profile is five times more
likely to live in a low space building than a high space one. The Asian population in Chicago is
consistently represented in all of the ranges personal space.
% Pop.
Total Population
High Density
Low Density
High Personal Space
Low Personal Space
Green
Purple
Orange
Pink

Total Pop.
% Area % White Total White % Black
Total Black
100
2,700,000 100%
32%
864000
32%
864000
60%
1,600,000
33%
31%
496000
24%
384000
15%
400,000
33%
45%
180000
40%
160000
30%
800,000
33%
55%
440000
29%
232000
40%
1,110,000
33%
26%
288600
27%
299700
15%
400,000
7%
57%
228000
25%
100000
30%
800,000
17%
24%
192000
17%
136000
7%
200,000
17%
53%
106000
35%
70000
3%
80,000
7%
17%
13600
45%
36000

% Hispanic Total Hispanic % Asian Total Asian % Other
Total Other
Total Population
21%
567000
4%
108000
11%
297000
High Density
28%
448000
4%
64000
13%
208000
Low Density
9%
36000
2%
8000
4%
16000
High Personal Space
7%
56000
5%
40000
4%
32000
Low Personal Space
32%
355200
3%
33300
13%
144300
Green
9%
36000
6%
24000
3%
12000
Purple
37%
296000
3%
24000
19%
152000
Orange
6%
12000
2%
4000
4%
8000
Pink
17%
13600
2%
1600
7%
5600
Table 5. Demographic profile of all spaces in Chicago. The demographic profiles for all
conventional high and low density areas, high and low personal space, and the four corners of the
legend are shown. Total population and each demographic group are shown for each
classification as both percent of total within that class and raw number of people.

Table 5 shows the full demographic profiles for the five race categories at each of the
following geographies: the entire city of Chicago, conventional high and low density, both high
and low personal space, and at each of the four classifications in the corners of the legend. The

76

demographic profiles of Chicago illustrates that not all places of classified together with
conventional population density measurements have the same type of demographic landscape.

5.5 Individual Representation Results
The purpose of the individual representation method is to capture the phenomenon of the
individual people living with the residential buildings distributed across the city. In this
approach, a dot density technique is utilized with dots being placed only in residential buildings.
The unit value is one dot per every person in the city, classified by race with hue. A Poisson
distribution determines the number of dots per building based on the expected frequency of
individuals from the personal space metric.

Figure 16. Poisson distribution. An example of a Poisson distribution where the expected
intensity of the output is a lambda equal to 5. This distribution allocates individuals to each
building in the city. Graph design by Nicholas A. Perdue, 2012.
77

The Poisson distribution allocates individual people in the census block back into the
residential buildings. A Poisson distribution is a discrete probability distribution that will extract
the probability of an event at a fixed interval based on a known average rate or expectancy
(Bailey & Gatrell, 1995). Calculating the expected personal space is in each building allows for
the creation of a series of visualizations showing the individual person confined to the building
level. The Poisson distribution forces a degree of variation within the buildings to avoid creating
a series of homogenous living spaces across the block area.
The final results of this method allocates the population totals from 16,885 census blocks
with a total of 2.7 million people into greater than 400,000 residential structures across the extent
of Chicago. The individual dots are symbolized with hue to represent one of the five census race
categories- white, black, Hispanic, Asian, or other. The chorodot allows for the attribute race to
be mapped while maintaining the size of the dot and the placement within the building.
The maps are created as pairs- one with the aerial image behind the dots and one of the
same scenes with the image removed. The initial image allows for the context of the space to be
seen, shows the built environment behind the individual, and illustrates the precision of the dot
placement. Removing the image and representing the individual in blank space allows for the
process of human settlement to be realized without the demarcations typically associated with
urban spaces.
Below are four examples from different community areas of Chicago. Each represents a
very different spatial structure and residential pattern. All the dot maps are represented at a scale
of 1:5,000.
The first set of maps is from Lakeview; a community area on the North Side of Chicago
classified as green in the personal space bivariate classification. Lakeview has a population of
78

107,291. Of that population, the percent white is 84%, black 4%, Asian 5%, Hispanic 4%, and
other 3%. The conventional density measurement is one person for every 812ft 2 of ground area.
The personal space index is 1,133ft 2 of residential space per person. When the image behind is
removed, a clear pattern of concentration can be seen in the right side of the image, indicating
many people living in high residential structures along the lakefront.

Figure 17. Individual representation map with image, Lakeview. The map shows the pattern
of residential settlement with the individual, represented by race, confined to the building in
which they live. The image behind the building gives context the space and the built environment
of the Lakeview community area. The vertical buildings along the lake have the highest
concentrations of people. Map design by Nicholas A. Perdue, 2012. Data provided by United
States Census Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS
department; background image provided by ArcGIS server.

79

Figure 18. Individual representation map without image, Lakeview. The map shows the
pattern of residential settlement with the individual, represented by race, confined to the building
in which they live. The removal of the background image reveals a pattern of residential
distribution in Lakeview without boundary demarcations or the context of the built environment.
This map allows for the pattern of how people reside in vertical sections of cities to stand alone.
Map design by Nicholas A. Perdue, 2012. Data provided by United States Census Bureau,
Chicago Metro Agency of Planning and by the City of Chicago GIS department.
The second set of maps is from Forest Glen; a community area in the Far North of
Chicago classified as orange in the personal space bivariate classification. Forest Glen has a
population of 19,974. Of that population the percent white is 86%, black 1%, Asian 9%,
Hispanic 3%, and other 1%. The conventional density measurement is one person for every
4,462ft2 of ground area. The personal space index is 760ft 2 of residential space per person. When
the image behind is removed, the spacious single family residential pattern and land use can be
seen.
80

Figure 19. Individual representation map with image, Forest Glen. The map shows the
pattern of residential settlement with the individual, represented by race, confined to the building
in which they live. The image behind the building gives context the space and the built
environment of the Forest Glen community area. Forest Glen shows a residential pattern of
suburban single family homes on large lots. Map design by Nicholas A. Perdue, 2012. Data
provided by United States Census Bureau, Chicago Metro Agency of Planning and by the City of
Chicago GIS department; background image provided by ArcGIS server.

81

Figure 20. Individual representation map without image, Forest Glen. The map shows the
pattern of residential settlement with the individual, represented by race, confined to the building
in which they live. The removal of the background image reveals a pattern of residential
distribution in Forest Glen without boundary demarcations or the context of the built
environment. This image allows for the pattern of how people reside in this space of Chicago to
stand alone. Map design by Nicholas A. Perdue, 2012. Data provided by United States Census
Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS department.
The third set of maps is from South Lawndale; a community area in the West Side of
Chicago classified as purple in the personal space bivariate classification. South Lawndale has a
population of 93,555. Of that population the percent, white is 3%, black 14%, Asian 0%,
Hispanic 81%, other and 2%. The conventional density measurement is one person for every
1,368ft2 of ground area. The personal space index is 239ft 2 of residential space per person. When
the image behind is removed, the crowded, tightly packed single family residential landscape of
the South Lawndale space can be seen clearly observed.
82

Figure 21. Individual representation map with image, South Lawndale. The map shows the
pattern of residential settlement with the individual, represented by race, confined to the building
in which they live. The image behind the building gives context the space and the built
environment of the South Lawndale community area. South Lawndale shows a residential
pattern of tightly concentrated single family homes on small lots. Map design by Nicholas A.
Perdue, 2012. Data provided by United States Census Bureau, Chicago Metro Agency of
Planning and by the City of Chicago GIS department; background image provided by ArcGIS
server.

83

Figure 22. Individual representation map without image, South Lawndale. The map shows
the pattern of residential settlement with the individual, represented by race, confined to the
building in which they live. The removal of the background image reveals a pattern of residential
distribution in South Lawndale without boundary demarcations or the context of the built
environment. This map allows for the pattern of how people reside in this space of Chicago to
stand alone. Map design by Nicholas A. Perdue, 2012. Data provided by United States Census
Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS department.

The final set of maps is from South Deering; a community area in the Far Southeast Side
Chicago classified as pink in the personal space bivariate classification. South Deering has a
population of 17,526. Of that population, the percent white is 4%, black 60%, Asian 0%,
Hispanic 29%, and other 7%. The conventional density measurement is one person for every
17,330sq2 of ground area. The personal space index is 358ft 2 of residential space per person.

84

When the image behind is removed, the crowded single family residential pattern with minimal
amount of the land use as residential can be observed.

Figure 23. Individual representation map with image, South Deering. The map shows the
pattern of residential settlement with the individual, represented by race, confined to the building
in which they live. The image behind the building gives context the space and the built
environment of the South Deering community area. South Deering shows a residential pattern of
tightly concentrated single family homes on small lots surrounded by a large amount of nonresidential land uses. Map design by Nicholas A. Perdue, 2012. Data provided by United States
Census Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS department;
background image provided by ArcGIS server.

85

Figure 24. Individual representation map without image, South Deering. The map shows the
pattern of residential settlement with the individual, represented by race, confined to the building
in which they live. The removal of the background image reveals a pattern of residential
distribution in South Deering without boundary demarcations or the context of the built
environment. This map allows for the pattern of how people reside in this space of Chicago to
stand alone. Map design by Nicholas A. Perdue, 2012. Data provided by United States Census
Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS department.

5.6 Summary
The results presented in this section illustrate many alternatives to conventional
population density. The personal space metric creates an index to articulate how much space
each person in the city has to themselves within the residential environment. This measurement
is critically important to understanding how people live within the built environment of the
contemporary city and the degree of crowdedness which is experienced by urban residents. The
86

measurement of personal space allows for geographers to capture the lived experience of urban
residents in Chicago at multiple scales with a singular metric. The bivariate classification of
personal space juxtaposed with conventional population density allows for demographic profiles
and racial breakdowns of place to be explored in insightful ways. Utilizing an enhanced measure
of urban space, key demographic variations can be observed in the different environments of the
city. The demographic profiles of different classifications of Chicago illustrate that the social
environment of the city is both a spatial and structural creation. Certain planning decisions and
mandates control the built environment of the city, the social landscape responds to the built
environment. The individual representation approach creates a series of new visualizations
showing the concentrations of humans in cities in unique ways. This approach shows how people
respond to the built environment by inhabiting the space provided by developers and residential
zoning regulations. The individual representation allows for a map reader to gain a clear
understanding of the residential environment without the demarcations typically associated with
urban spaces.

87

Chapter 6
DISCUSSION

6.1 Overview and Goals
The purpose of this section is to discuss how effectively the various methods and
subsequent results presented in this thesis answer the three core research questions:
1. What are the limitations of traditional population density?
2. How do these limitations confuse our understanding of modern urban settlements?
3. How can emerging geographic analytics help address these limitations and generate a
more complete and precise understanding of urban population distributions?
The primary objective of this research is to identify the limitations of current methods of
representing urban population distributions and to implement new approaches to map and
understand populations in contemporary cities. The results and the analysis of the alternative
cartographic approaches contributes to the knowledge of urban population distributions by
illustrating how conventional methods fail to accurately characterize the vertical structure of
contemporary cities. The personal space metric demonstrates the variability in the demographic
landscape of cities when the vertical residential environment is accounted for in spatial analytics.
Comparing the demographic profiles of conventional and enhanced measurements of population
space, critical patterns of racial distribution, segregation, and urban residential conditions
emerge. The creation of alternative representations of the urban space serves to eliminate many
of the errors and misrepresentations of conventional techniques.

6.2 Personal Space Metric
The primary limitation of conventional density metrics is that it limits the analysis of
population dynamics to a two-dimensional space. Conventional density fails to reflect the true
88

manner in which people are distributed across the modern urban space. People live in vertical
structures across the developed urban landscape; the personal space measurement aims at
accounting for this vertical residential environment. An aerial image of North Side of Chicago
illustrates the vertical residential environment of conventionally ‘dense’ urban spaces; many
people living in the neighborhood space but structured vertically.

Figure 25. Vertical residential environment in Chicago. The view from the John Hancock
tower looking north along the Lake Michigan coast, including the community areas of Lakeview,
Lincoln Park, and Uptown. The vertical residential structure of the city is evident in the
apartment and condominium structures of this part of the city. Photo taken by Nicholas A
Perdue, 2011.

Mix use buildings present a challenge in the modeling of residential environments in
contemporary urban spaces. Mixed-use development refers to zoning of a building or a group of

89

buildings with more than a singular purpose- to be not solely residential or commercial space.
Mix use developments are designed to spatially cluster employment, housing, commercial, and
recreational activities together in a centralized urban space (Duaney, 2002).

Figure 26. Mix Use Building, Lincoln Park Chicago. A mix use building in the community
area of Lincoln Park presents challenges in modeling due do to the mixture of commercial and
residential spaces within the building. Mix use buildings are becoming more popular in the
gentrification of the old urban core and the redevelopment of the space. Photo taken by Nicholas
A Perdue, 2011.

A collection of similar style mix use buildings, or a New Urban District, is preferably
spatially close to a public transit node and is often the result of redeveloped urban areas or as part
of a planned town center (Lees, 2008). In specific zoning terms, mix use development refers to
some combination of residential, industrial, office, commercial, and institutional land uses
(Duaney, 2002). Mix use buildings are common features in the redeveloped, recentralized urban
space and need to be modeled carefully to fully understand the amounts and types of living

90

spaces in contemporary cities. In the personal space metric, 60% of the space within mix use
buildings is assigned as residential and 40% non-residential.
Dividing mix use buildings into residential and non residential in this manner adds a
degree of uncertainty to the variable livable space. A study to evaluate the interior of multiple
mix use buildings in different locations across Chicago could have been employed in order to
reduce the uncertainty. It is suggested however, in urban development simulations that the
degree of variation in mix use buildings is inconsistent and continuously changing as occupancy
and ownership changes. One of the advantageous of mix use development is the ability to change
the building typology without the hassle of planning and code changes. For this reason, models
have relied on the 60% estimation as the most appropriate metric (Waddell, 2003).
The personal space metric reduces the limitations of conventional density measurements.
Analytical research on the spatial distribution of people often conceptualizes individuals based
on standardized units. People in contemporary cities do not populate only units of ground area.
Rather, individuals are distributed in discrete structures built vertically across the city.
Conventional population density aims to eliminate the errors and misrepresentations attributed to
the size of an enumeration unit; it does not eliminate the errors associated with the vertical space
problem. The vertical space problem can be described as simply the assumption of homogenous
vertical extents within the enumeration unit in conventional population analytics. For populations
structured in a uniform manner, conventional population density is a sufficient measure;
measuring the number of corn plants per square acre is an adequate utilization of population
density. Measuring concentrations of people in contemporary cities requires a more complex
metric. The personal space metric meets the complexity of the contemporary urban structure and
form.

91

6.3 Classification of Personal Space
Juxtaposing the new personal space measurement and the conventional population
density measurement in a bivariate map establishes the distinction between crowdedness and
density in the city space. All conventional high density spaces are not the same; there is a range
of residential environments within a high density urban space. While it is important to
understand how many people live within certain area for planning and resource purposes, density
is not a measurement of crowdedness or of living conditions. The classification system illustrates
both conventional density and personal space on a grid where the interactions between high,
medium, and low values for both variables can be observed. The four corners of the legend
design represent the spaces where the highs and lows of the two metrics interact.
Each of the four classifications in the corners of the legend represents a typology of the
urban form. The green classification characterizes the vertical urban landscape where high-rise
buildings and tightly packed vertical houses are the prominent features. The purple classification
indicates short, tightly packed small urban homes and apartments on small properties. The
orange classification denotes two different types of landscapes. The first is a suburban form with
large houses on large plots of land; the second, the transformation of commercial or industrial to
redeveloped residential spaces. The pink classification signifies a fragmented landscape with
small housing complexes surrounded by commercial and industrial land uses. The four
classifications have a unique social, structural, and spatial composition revealing some of the
underlying complexity in urban population analytics.

92

6.4 Demographic Profiles of Chicago
The comparison of demographic profiles between different high and low density spaces
in Chicago reveal how places with identical classifications in a conventional measure differ
greatly in structure and social pattern. The demographic profiles allow for a glimpse into how
Chicago is physically developed and the extent to which racial segregation is distributed across
the built environment and the residential structures of the city.
The green classification exemplifies the necessity of the personal space metric. The green
spaces are mostly concentrated in the North Side along the Lake Michigan shoreline and
represent the places in Chicago dominated by the spacious vertical residential environment. The
population in the green class is disproportionately higher in percent white than conventional high
density blocks and all of Chicago in general. The percent white increases at the expense of other
races, most notably the Hispanic population.

93

Figure 27. Conventional high density comparison. The two classified areas in this map show
the range of personal space within the Census blocks classified as high density spaces by
traditional population density methods. The two spaces, though both high density, represent very
different spaces both in terms of the built environments and the demographic profiles. Map
design by Nicholas A. Perdue, 2012. Data provided by United States Census Bureau, Chicago
Metro Agency of Planning and by the City of Chicago GIS department.
The purple classification also represents a conventional high density space, but with low
amounts of personal space. In the purple class, the percent white decreases dramatically as the
percent Hispanic notably rises. The inclusion of personal space significantly changes the
demographic profiles of the conventional high density designation. For example, the statement
that 28% of the population in high density spaces of Chicago is Hispanic is misleading.
Accounting for the vertical space in the classification system, the percent Hispanic ranges from
9% in the spacious to 37% in the crowded densely populated spaces.
94

The type of buildings characterizing the range of conventional high density spaces is
equally as dramatic. The green classification consists of high rise apartments and large, free
standing single family homes or duplexes. There is a tight concentration of buildings on the
street with small yards and little setback between houses. The buildings, however, typically
contain a high amount of vertical residential space. According to Real Estate Website Zillow,
prices in Lakeview show a 1,000ft 2 apartment ranging from $2,000-$6,000 a month depending
on proximity to the lake, age and type of building, distance to certain commercial or communal
areas, view shed, and floor. Buying property in Lakeview ranges from $500,000 to multiple
millions depending on a range of variables. The green classification in Chicago is an exclusive,
wealthy, spacious, and predominately white city space. Lakeview, with the highest concentration
of high rise residential structures in the city, is 84% percent white; figures 28 and 30 shows the
residential landscape of the Lakeview community area.
The buildings in the purple classification also exhibit a tight concentrated on the surface
of the earth, but contain little in terms of vertical residential space. In the purple spaces, the
measurement of high density equates to a crowded residential space. Real estate prices in South
Lawndale, for a 1,000ft 2 apartment range from $400-$700; buying property will typically cost
$50,000 to $75,000; figures 29 and 31 shows the residential landscape of South Lawndale

95

Figure 28. High density, high personal space apartment. This high rise apartment, located in
Lakeview overlooking Lake Michigan is typical of the high density, high personal space multioccupant building. Photo taken by Nicholas A Perdue, 2011.

Figure 29. High density, low personal space apartment. This three story apartment, located in
South Lawndale is characteristic of the high density, low personal space multi-occupant building.
Photo taken by Nicholas A Perdue, 2011.
96

Figure 30. High density, high personal space house. The single family free standing
brownstones present in the Lincoln Park community area are consentient with many of the high
density, high personal space single-family buildings in Chicago. The houses are tightly packed
with a high vertical extent. Photo taken by Nicholas A Perdue, 2011.

Figure 31. High density, low personal space house. The single family home that characterizes
the high density low personal spaces. This house, located in South Lawndale exhibits a built
environment of small houses tightly packed together. Photo taken by Nicholas A Perdue, 2011.
97

The purple classification is a predominately segregated space in the city of Chicago. The
percentage white in the purple class is less than half that of the green class. There are 30,000
more white people living in the 7% of the city classified as green than there are in the 17% of
Chicago classified as purple. Conversely, there are 265,000 fewer Hispanic in the green spaces
than in the purple. The areas classified as purple are crowded, lower class, and predominately
segregated spaces in the city of Chicago.
The demographic profiles of the conventional low density spaces of Chicago show a
similar pattern in demographic variability with the introduction of the personal space metric. In
the orange classification, the percent white is 53%; in the pink classification, the percent white is
28%. The decrease in white in the pink classification is offset by an increase in percent black and
Hispanic. Over 60% of the white individuals living in the conventional low density spaces of
Chicago live in the orange classification. The most spacious low density blocks, visualized with
the orange classification, represent two unique types of urban space in Chicago.
The first and most abundant of the orange classification is the suburban development
along the perimeter of the city. In places such as Forest Glen, spacious lawns with large house
dominate the landscape. The white race accounts for 86% of the residents in Forest Glen. The
suburban regions classified as orange in the Far North and the Far Southwest of the city were
among the final parts of the city to be annexed, were developed around the streetcar, and have
maintained a suburban structure since becoming part of Chicago. Property in Forest Glen ranges
from $250,000- $1,000,000 depending on size and proximity to one of the two golf courses in
the community area. Figure 33 is an example of the type of structure common to Forest Glen

98

Figure 32. Conventional low density blocks comparison. The two classified areas in this map
show the range of personal space within the Census blocks classified as low density spaces by
traditional population density methods. The two spaces, though both low density, represent very
different spaces both in terms of the built environments and the demographic profiles. Map
design by Nicholas A. Perdue, 2012. Data provided by United States Census Bureau, Chicago
Metro Agency of Planning and by the City of Chicago GIS department.
The second type of orange space is the recently redeveloped areas of the city, especially
the neighborhood spaces in the Near South Side. Redevelopment of the Near South Side with
modern mix use buildings began in the late 1990’s as property values in the North Side of
Chicago grew rapidly. The redevelopment plan sought to emphasize the museum and stadium
districts just south of downtown and utilize the industrial infrastructure converting many of the
old warehouses on Printer’s Row into modern mix use lofts and studio space (Stelle, 2011). The
wave of redevelopment was brought on by an artistic class seeking central locations, cultural
99

amenities, and access to cheap studio spaces in the mid 2000’s (Ley 2003). This process of
redevelopment is driving the landscape of the South Side to match the structure of the North
Side. As a result, conventional measures of population density are low as many buildings have
yet to be or are in the process of redevelopment. The residential environment within the
buildings, however, is spacious- leading to a classification that matches the suburban space. As
the process of redevelopment continues, the population density of the Near South Side should
increase and move towards the green classification. Figure 35 is an example of the mix use
redeveloped residential structures common in the orange classified Near South spaces of
Chicago.
The pink classification exhibits a different dynamic in Chicago; spaces where small
residential structures are built in a mostly non-residential area. The pink areas are some of the
least expensive and least occupied spaces in Chicago, often a fragmented landscape of residential
buildings along industrial and transpiration corridors. Less than 3% of the population of Chicago
resides in the pink class. The desertion of the steel industry in South Deering left a fragmented
post-industrial landscape where small pockets of residential structures are separated from each
other by large corridors of abandoned industrial buildings and waste (Rosen, 1998). South
Deering is 60% black and 30% Hispanic. Property in South Deering ranges from $20,000 to
$40,000 depending on location within the community area. Figures 34 and 36 are examples of
the residential landscape in South Deering.
The personal space metric captures the residential environment of the contemporary city.
Spaces identical according to conventional analytics are clearly very different pieces of the urban
geography. The personal space measurement fundamentally changes the conceptualization of

100

what high density and low density urban places are by examining residential patterns in
conjecture with the built environment.

101

Figure 33. Low density, high personal space house. The single family home that characterize
the low density high personal spaces. This house, located in Forest Glen exhibits a suburban built
environment of large houses on large lots. Photo taken by Nicholas A Perdue, 2011.

Figure 34. Low density, low personal space house. The single family home that is typical of
the low density low personal space parts of Chicago. This house, located in South Deering is a
small house on a relatively large lot. The construction of houses in these areas is often close to
industrial and transportation corridors creating fragmented spaces. Photo taken by Nicholas A
Perdue, 2011.

102

Figure 35. Low density, high personal space loft. A recently gentrified mix use building in the
printer’s row neighborhood of the Near South Side. This part of Chicago is undergoing a shift in
population and types of residential structures in the past 10 years. Photo taken by Nicholas A
Perdue, 2011.

Figure 36. Low density, low personal space house. Trailer Parks are a type of residential
structure that is low density but crowded residential spaces. These community, in South Deering
exhibits the low density, low space classifications of residential environments close to an
industrial corridor. Photo taken by Nicholas A Perdue, 2011.
103

6.5 Individual Representation
The dot density maps showing individual representations of people in Chicago allocate
totals from census block totals into individual buildings. The Poisson distribution allows for
variations of personal space within the buildings and prevents a homogenous representation. The
imposed variation replicates the lived experience inside of buildings- one house of a certain size
may have three individuals living within; an identical house next door may have four. The
Poisson method captures the micro-scale variations of urban residency. The limitation introduced
with the method, however, is that the allocation of the individual is random based on the
expected intensity of people and the amount of space in the building.
The dot density maps accurately characterize the residential pattern of individual
distribution within buildings and match the residential patterns of the census block. The measure
is precise in that individuals are represented in only the residential buildings within the census
space. The limitation lies in that micro-variations may exist may not be fully accounted for by
the metric.
Figure 37 shows an accurate portrayal of the residential pattern for a section of Hyde
Park, Chicago. The background image provides the context of the built environment of the
neighborhood and shows a texture of vertical space. The two horizontal buildings in the center of
the image are an eleven story condominium structure that really drives the conventional
population density metric within the boundary area. This census block group is a high density
area with conventional measurements, mostly as a result of the concentrations within two
buildings that differ greatly from the other features in the image. Figure 38 provides a look at the
pattern of individuals in space without boundaries. With all features, boundaries, and identifying
place marks removed from the map, an entirely different sense of population is realized.
104

Figure 37. Dot density map of Hyde Park, with image. The map shows that the residential
pattern in Hyde Park on the South Side of Chicago. The image behind the classification provides
context into the structure of the built residential environment and geographical features. Map
design by Nicholas A. Perdue, 2012. Data provided by United States Census Bureau, Chicago
Metro Agency of Planning and by the City of Chicago GIS department. Background image
provided by ArcGIS server.

105

Figure 38. Dot density map of Hyde Park, without image. The map shows that the residential
pattern in Hyde Park on the South Side of Chicago. The removal of the image behind the
classification shows that the residential pattern is an organic response to the spaces available in
the built residential structure and that the distribution of people is not homogenous across urban
spaces. Map design by Nicholas A. Perdue, 2012. Data provided by United States Census
Bureau, Chicago Metro Agency of Planning and by the City of Chicago GIS department.

Individuals fill the space, both planar and vertical as an organic response to the built
environment around them. The infrastructure of the urban space provides a spatial configuration
of individual residential behavior that operates not on the boundaries of neighborhoods or census
blocks, but rather on the functionality of the residential environment.
106

Another limitation of note is the model may inaccurately integrate racial representations
together. The pink dots in the map above represent white individuals in the block; the blue dots
black individuals. Although the space appears to be integrated, a clear separation in the microprocesses of neighborhood dynamics may exist that segregates the space and causes all of one
race to be in one section of buildings and all of another race in a separate set of buildings set
apart by a defined geographical or social feature. The model used to create this series of maps
will not account for that type of dynamic. The individual representation maps do, however,
provide a unique look at how the people occupies the built environment of the urban space, show
meaningful patterns, and provides significant insight into urban residential arrangement.

6.6 Contribution to Geography
This research contributes to the fields of cartography, population analysis, and urban
geography in three distinct ways. The creation of the personal space metric matches the form of
contemporary cities. This metric generates greater insight into the complex relationship of people
in urban space. The personal space metric creates a new measure in which to evaluate urban
population concentrations. The distinction between conventional density and living space allows
for demographic trends and indicators of developmental policy to be identified in the urban
landscape. Through the design of personal space and classification in conjuncture with
conventional population density, urban analytics of city populations are applicable at multiple
scales and across multiple geographies.
Many representations of the urban space in the field of geography present a simplified
and inaccurate portrayal of the manner in which individuals settle in the residential spaces of the
city. Many mapping approaches rely on conventional population density as a primary input

107

variable. In urban spaces, the idea of crowdedness is commonly linked to a high density
environment. Contemporary urban spaces are prominent displays of vertical residential structures
that accommodate many individual people in small geographic space. A large number of people
living within a surface unit is not an indication of crowdedness. The vertical structure of the city
allows for many people to share a common ground space in the city yet have ample personal
space. In the generalized representations of conventional density, very different lived
experiences of space are represented with a similar density values; a result of the reliance on
boundaries and units and a failure to conceptualize the vertical spaces of urban environments.
The personal space metric accounts for the vertical residential environment of cities and
effectively differentiates between different types of conventional high and low density spaces.
Conventional representations of the urban space, in many cases, present a simplified and
inaccurate portrayal of the manner in which individuals settle in the residential spaces of the city,
simplifying space as a high density or low density space without consideration to how the
individual person occupies the individual structures. Contemporary urban spaces are filled with
numerous residential structures of varying height within a neighborhood, accommodating many
individual residents, with a range of personal spaces. The modeling of the third dimension of
space, the process of deconstructing the invisible vertical boundary in urban analytics, is
essential in the creation of conceptual models to understand the experiences of crowdedness or
spaciousness often mistaken with density.
Urban populations should be analyzed as a function of populated residential buildings,
not as a function of totals within a defined boundary. From images and experiences in cities, the
urban space is clearly not a homogenous landscape across the extent of administrative units. The
removal of the urban boundaries in analysis and representation is a crucial step in creating a

108

richer understanding of the distribution of people across the backdrop of the urban space. The
spatial structure of individuals is evident in the dot density approach designed to be accurate in
representation and precise in location. The individual representation model is designed to reduce
the uncertainty and generalization in the cartographic symbology typically associated with dot
density maps. This approach eliminates the demarcations of boundary units and borders and uses
only residential structures of the urban space. Each individual is represented as a single entity,
creating a representation that shows how individuals organize themselves into the environment
around them. The primary purpose of the individual representation approach is to capture a fine
scale representation of urban residents in space.
As cities become increasingly more recentralized and intentional developmental
strategies are employed to create more vertical and self contained urban structures, the
functionality of modeling personal space becomes critical. The personal space metric leads to an
important discussion about the nature of the changing urban structure and the differences in
places of similar conventional population density across the landscape of cities. The personal
space metric and the demographic breakdowns of the improved classification system illustrate
not all urban spaces are the same, even when represented similarly under conventional
techniques. The green spaces, for example, have a different type of landscape than the purple
spaces. The individual representation maps and qualitative analysis of the green space indicate
that the residential spaces of the green class classification are occupied by mainly affluent white
individuals. In contrast, the purple space shows a congested housing environment with little to no
variation in the vertical landscape. Neither the purple nor the green spaces depict the average or
prototypical high density development. Both places are high density spaces that are common in
all cities. The primary distinction is the vertical extent of the built environment. It is clear that

109

the developed vertical residential space of cities represent a unique urban condition from a
structural and social context.
In the context of policy and planning decisions, the framework of personal space and
improved classifications allow for the processes of urban redevelopment to be conceptualized
with greater depth. The process of redevelopment is aimed at attracting a more affluent set of
residents, to concentrate commercial investment, to increase homeownership and the local tax
base (Smith, 1996). Redevelopment is the process of revitalizing struggling and decaying
neighborhoods in the urban core of cities. Ideally, though rarely the case, policy and
development initiatives view redevelopment as a mechanism to bring about socially mixed, less
segregated, sustainable, and more livable communities (Lees, 2008). It is suggested a revitalized
neighborhood will cultivate social mixing, social capital, and social unity in the community
space. The process of urban redevelopment allows for affluent property owners to relocate to the
city for the cultural benefits for urban life with the relative security of suburbia (Lees, 2008). In
actuality, redevelopment commonly has negative repercussions for the larger community.
Sociocultural diversity is a leitmotif in the new taste for the central city housing and
neighborhood. One of the great amenities of dense city living is exposure to such social
and cultural diversity as ethnicity. The urban ambiance of diversity is a continual source
of simulation and renewal and a reminder of the cultural relativity of one’s own style of
life. It is said to be a relief from the subculture sameness and ‘boredom’ of many
suburban communities (Allen, 1984, pp 31-32).
Redevelopment of the urban core typically involves a movement of the affluent into an
urban space accompanied by the exodus of the lower class into another space, a process
generally referred to as gentrification. Many studies have found that redevelopment and
gentrification are followed by declining levels of social mixing and ethnic diversity; redeveloped
neighborhood spaces typically become exclusive- lacking transference between socioeconomic
groups, lacking cohesion between the new buildings and the old (Lees, 2008). In the context of
110

Chicago, the redeveloped urban spaces exhibit three characteristics that are captured by the green
classification- high personal spaces in a vertical environment, exclusiveness, and lowered rates
of ethnic diversity. The green classification captures and represents the new urban spaces of
Chicago, the places that have undergone this crucial spatial and structural change.
The distinction between the green and the purple classification can also be clearly seen in
the amount of the space in these census blocks is residential versus the amount that is nonresidential. Both classifications have similar amounts of people per unit of ground area but the
green, as has been demonstrated, has a vertical environment where people reside whereas in the
purple people reside much closer to the ground. The non residential space in the green places,
consisting of 52% of the ground space shows a much different landscape than the 24% nonresidential in the purple class. There is a strong importance of the legislative and planning
process in the development of high density urban areas. The intensification of people, from a
developmental standpoint, is aimed at generating economic situation not seen in lower density
spaces. Intensification of people creates unique, niche markets and elastic demand that can be
captured by the diversification of business and integration of commercial and dense residential in
a singular place.
Planning and development policy draws a distinction between a high density urban core
and a high density urban residential (Vojnovic, 2000). The green classification, by showing an
intensification of people as well as a majority of the land designated as non-residential meets the
requirements of a strong urban core. The diversification of the urban core into a mixture of high
rise residential and niche business creates a flexible and vibrant economic space. The strong
urban core is a perfect mixture of economic, social, and cultural assets to support and
accommodate the affluent. The green spaces represent the unique residential and economic

111

advantages of the urban milieu. The non-residential spaces it the green areas are filled with
boutique shops, restaurants, and vibrant nightlife activities. The non residential space of the
purple classification is mostly roads and generic businesses. Just as there is a distinction in the
type of residential structure between the green and purple class, there is a notable distinction in
the opportunities and enterprises of the non-residential. The process of redevelopment allows for
the transformation of the spaces to the desirable green class.
In Chicago, redevelopment has historically been clustered primarily in the North Side.
Community areas such as Lakeview, Lincoln Park, Uptown, and the Near North Side were in an
advantageous structural and spatial position for the redevelopment of the community space- large
quantities of older brownstone housing, adjacency to Lake Michigan, and short commuting via
mass transit to the downtown Loop. In the late 1950’s the first redevelopment program in city of
Chicago was initiated in Lincoln Park, ensuing in extensive housing demolition and
displacement in the Old Town neighborhood in the southeast portion of the community area,
(Bennett, 1990). Within a few years, most of Old Town’s Polish community had moved towards
the suburbs and many of the old buildings in the neighborhood bought and restored by an
emerging educated, politically active middle-class actively seeking urban culture and amenities
(Bennett, 1990). Between 1970 and 1990, the majority of the North Side of Chicago was
redeveloped by an urban upper middle class. Different types of redevelopment strategies have
spread to other parts of Chicago to meets the demands of different groups of people, notably in
Wicker Park and recently in Logan’s Square. The North Side, however, in the development of
spaces for an affluent subset of people has relied on large-scale developers to redesign of
residential space and build a vertical landscape (Bennett, 1990).

112

The green classification represents a part of Chicago with a quality of desirability and
aesthetics in the urban structure. The vertical environment of Chicago is largely dependent on the
proximity to the lakeshore and the majority of the high rise residential structures are distributed
either along Lake Shore Drive or within a close distance. Vertical buildings, from a planning and
construction standpoint, cost considerably more in terms of investment, capital, and materials
than the same size parcel of land built with only a house or small apartment complex. A high rise
building cost more per square foot of floor space due to the cost of vertical transportation of
materials, site setup and appropriate infrastructure, and the increased pay and safely precautions
for workers operating at heights (Blackman and Picken, 2010). Lots zoned having a high FAR in
planning demand complex vertical structures that can maximize the available square footage of
floor space allotted by legislative action. The distribution of FAR is closely aligned to the
distribution of high rise structures across Chicago and of the green classification of residential
space. Development policy and zoning of FAR has an enormous impact on the residential
environment of city spaces and the resulting social atmosphere of the space. This construction is
understood and often employed in the fields of urban planning, developmental policy, and civil
engineering. The personal space metric is a step towards matching that conception in geography
and cartography.
The cost of developing the vertical residential environment translates to large scale
development and creates a region of exclusiveness for the wealthy in the city. The high rise
building contributes to the skyline of the city, a physical representation of the city’s influence,
economy, status, and desirability. Skylines are depicted in art and pop culture and are unique
identifiers of culture and social capital for those occupying the space (Spreiregen, 1965). Due to

113

this, the tall buildings in the cities will often be unique designs found on the land with the highest
desirability, visibility, and aesthetic qualities.
A second type of building also fits into the green classification, a structure common to
the North Side of Chicago where the bulk of the green classified blocks are located. The North
Side is characterized a skyline of high-rise apartments and condominium complexes along the
lakeshore. The western sections of the North Side community areas are characterized by high
end commercial and entertainments districts mixed with two to four-story single family
brownstone neighborhoods. The brownstone design typically refers to a free-standing house with
little or no yard, originally built as a single family house (Channick, 2008). The brownstone
houses and apartments are spacious, expensive, and exclusive residential structures in Chicago.
Lincoln Park, for example, is one of the wealthiest in the country with the average single family
house is priced around $1,000,000 (Woolsey, 2007). The green classification serves as an
indicator of redevelopment within the urban core simply from the juxtaposition of conventional
density and the personal space metric.
The transition of one classification to another over time is an important artifact of the
personal space measurement. The metric allows for the observation of transition of city space in
conjuncture with policy and development. The transition of a purple class to a green class with a
planned redevelopment project will be revealed with a temporal analysis of personal space. The
transition would encompass the developmental policy of shifting from a high density residential
to a high density urban core. The process includes a shift in the residential structure as well as
neighborhood composition. Cabrini-Green serves as an example of this type of transition that has
taken place in Chicago over the past 20 years.

114

Cabrini-Green is a neighborhood in Chicago and serves as an example of the
redevelopment process in city space. Cabrini-Green exemplifies a space where urban analytics
would serve to better understand the processes of redevelopment. At one time, Cabrini-Green
was a public housing project with rampant crime, poor infrastructure, and severe segregation
(Bennett, 1998). Beginning in the late 1990’s the public housing units were demolished and the
existing residents pushed out.

Figure 39. Cabrini-Green redevelopment. The redevelopment of Cabrini-Green is illustrated
by the last remaining public housing building in the background and the newly developed
residential space in the foreground. The transition of the city space can be observed in the
personal space metric but is not detected in conventional approaches. Photo taken by Lawrence
J. Vale, 2011. Photo courtesy of Archinect News.

115

The space was redeveloped and reoccupied by a very different group of people with very
different expectations about spatial amenities (Bennett, 1998). In the process of fifteen years, the
form and demographic composition of the space was completely changed. Through the process,
the conventional population density remained similar; both time periods would be classified as
high density with conventional metrics. The amount of space within the buildings changed
dramatically however. Whereas once Cabrini-Green would have been classified as a purple
space, it is now a part of the green classification. Figure 40 shows the composition of transition
in Cabrini-Green. In the background is the William Green Homes high-rise, a crowded public
housing structure that would have been classified as purple. The building was demolished in
2011 and was the final structure from the former Cabrini-Green housing project. In the
foreground the Old Town Village West homes, a 2009 development as part of the reclamation of
the Cabrini-Green space. Conventional techniques to population analysis show these two types
of places as equally dense areas. Enhanced metrics show the critical structural change of the
space. Since 2000, the Chicago Housing Authority has demolished over 100 subsidized housing
structures including all structures in Cabrini-Green. (Channick, 2008). The structures have been
replaced with three to four story mix use inclusive developments. The townhouse development
that has replaced the high rise projects has the same conventional density as the previous space
but people are organized in a different structure. Rather than a concentrated residential space
surrounded by the parking lots and open space, the buildings are integrated with boutique
commercial and restaurants. The public space in the Cabrini-Green area includes breweries,
upscale furniture retailers, and antiques stores. The transition into the green classification is
exhibited in both a spacious private space and a culturally active public space. The transition
demographic landscape from predominantly black to predominately white has followed the

116

structural shift. The enhanced analytics of the personal space metric captures this changing urban
landscape, conventional metrics do not.
Redevelopment of the urban core is a process that has and will continue to reshape the
structural and socioeconomic landscape of all cities. Urban analytics with enhanced metrics
allows for the study of complex and robust urban datasets and the comparisons of people and
space both across geographies and through time. The evolution of a single space over time
indicates the role development and policy has on the urban environment and the people within
the space. The transition of purple to green as exhibited in the Cabrini-Green example is a
function of the development policy. The enhanced personal space metric captures this change.
This allows for the study of not only what changes has happened but as a predictor for where
change is likely to occur in the future.
The movement of classification can also be observed in the orange class and former
industrial space is converted into residential space along Printer’s Row in the Near South Side.
Initially this space is non-residential. Through development initiative, the abandoned industrial
infrastructure is converted to mix use developments and luxury residential spaces. The space
falls into the orange class as there are few developments in the census block resulting in a low
population density. There are high amounts of personal space in the private spaces in the
buildings. This results in an orange classification that is commonly found in the suburban
structure. Continued development of this space will result in an increased population density
with housing structure high in personal space. The transition in this case is non-residential
through the orange classification and into the green.
Another transition that will be evident with the enhanced analysis is the abandonment of
formerly densely populated residential neighborhoods. A space that was once purple starts to

117

lose population without the space being redeveloped or reoccupied. The small housing with little
personal space remains but the neighborhood transitions from high density to low density. One
can imagine this would be the type of transition that would dominate the post industrial rust belt
cities of Detroit, Cleveland, and Milwaukee. This type of transitional analysis would provide a
richer foundation from which to evaluate the health of a cities population and living conditions.
Another valuable function of the enhanced metrics is the comparisons across
geographies, either from two locations in the same city or between different cities. Contrasting
different cities allows for direct analytics of how local policy and planning decisions reshape the
urban space. Urban analytics of the complex relationship of people and space allows for
quantitative analysis of movements and developments and an enhanced understanding of micro
to regional scale processes at work in cities. The deconstruction of conventional metrics can
usher in analytical measures for the characterization of aesthetics and desirability of spaces
within a city, can cohesively link how policy changes the urban landscape, and can point to
spaces where future change is likely to occur.
The personal space metric and the urban analytics derived by the interactions of spatial,
social, and structural characteristics are not exclusive to Chicago. The availability of data, the
unique historical processes, and the city geography makes Chicago an ideal space to explore the
concepts and metrics presented in this thesis. Similar analysis could be performed in any urban
space where the vertical environment and processes of development creates an urban landscape
where conventional measure will fall short of accurately characterizing the contemporary city
space. The personal space metric provides an alternative perspective on how people are
distributed in urban spaces with a relatively simple approach of spatial allocation based on the
zoning and the vertical extent of buildings within a city. Urban planners conceptualize the urban

118

space as a complex system of people, space, and structure; the conventional approaches in
geography and cartography do not. The failure to capture this complex relationship in geographic
discussions decreases our ability to understand the rapidly urbanizing populations of the world.
Chicago serves as an example of how alternative approaches and advanced urban analytics can
increase the depth of knowledge about how people are distributed across contemporary cities.

119

Chapter 7
CONCLUSION

7.1 Conclusion
This thesis introduces the vertical space problem as the primary limitation of
conventional population density in contemporary cities. Exploring urban population distribution
through the lens of Chicago with the personal space metric allows for a more complete picture of
contemporary cities. Urban analytics of concentrations of populations and demographic trends
reveal new patterns in the landscape of contemporary urban spaces. Conventional approaches
such as population density fail to characterize the lived experience of people within the
residential structures of buildings. Modeling the three dimensional residential space creates new
measurement crowdedness within the built environment of urban spaces.
The personal space metric developed in this thesis aims to deconstruct conventional
density, focusing on the relationship of the individual person and the residential space across the
extent of the city. When personal space is juxtaposed with conventional measures, limitations of
approaches which rely on administrative geographies and two dimensional spaces are exposed. A
dasymetric approach allocates individual people to three dimensional residential structures
allows for spatial, social, and structural processes in modern cities to be analyzed and
represented in unique ways. The personal space model is designed to be repeatable across
different cities but robust enough to capture the complexity of form in contemporary cities.
Demographic profiles contrasting conventional and new measurements capture how the
social landscape of cities change when the structural differences of residential spaces are
accounted for. The personal space metric also serves as an indicator of redevelopment policy in
the urban space. Different types of buildings frequent in redevelopment strategies exhibit
120

characteristics captured by the personal space measurement. Conventional measurements fail to
explain the differences between building typology in contemporary urban spaces. The individual
representation approach allows for the residential behavior of people in different spaces and built
environments to be explored with more depth than conventional techniques. The analysis of the
improved population measures results in a greater understanding of the complex patterns of
people in city spaces and improved insight into processes of redevelopment. Urban analytics of
the improved metrics of space in contemporary cities creates links between developmental policy
and the perceptions of residential space. Original metrics and representations allow for the
complex relationships of people in city space conceptualized by urban planners and experienced
by urban residents to be understood by geographers at multiple scales.

121

REFERENCES

122

REFERENCES

Adobe Flash CS5.5 (Version Creative Suite 5.5). (2011). San Jose, California: Adobe
Systems, Inc.
Adobe Illustrator CS5.5 (Version Creative Suite 5.5). (2011). San Jose, California: Adobe
Systems, Inc.
Adobe Photoshop CS5.5 (Version Creative Suite 5.5). (2011). San Jose, California: Adobe
Systems, Inc.
Aina, Tade Akin. (1992). Sustainable Cities: Meeting Needs, Reducing Resource Use and
Recycling, Reuse and Reclamation. International Institute for Environment and
Development.
Allen, Irving L. (1984). The ideology of dense neighborhood redevelopment. Gentrification,
Displacement and Neighborhood Revitalization. Eds. J J Palen, B London. State
University of New York Press, Albany, NY. 27-43.
American Fact Finder Help: Race and Ethnicity. United States Census Bureau. Retrieved
11/8/2011.
Anderson, J.F., E.E. Hardy, J.T. Roach, and R.E. Witmer. (1976). A land use and land cover
classification system for use with remote sensor data, U.S. Geological Survey
Professional Paper 964.
Anderson, M. et al. (1999). Sampling-based adjustment of the 2000 census-A balanced
perspective. Jurimetrics 40. 341-53.
Anderson, M., and S. E Fienberg (1999) Who Counts? The politics of Census-taking in
Contemporary America. Society 34(3), 19–26.
ArcMap 10. (2010). Redlands, California: Environmental Systems Research Institute, Inc.
Beckmann, Martin J. (1969). On the distribution of urban rent and residential density. Journal of
Economic Theory 1(1), 60-67.
Bertin, J. (1983). Semiology of Graphics. University of Wisconsin Press.
Bailey, T.C. and Gatrell, A. C. (1995). Interactive Spatial Data Analysis. John Wiley and Sons,
New York, NY.
Bennett, Larry. (1998). Do we really wish to live in a communitarian city? communitarian
thinking and the redevelopment of Chicago’s Cabrini-Green public housing complex.
Journal of Urban Affairs 20(2). 99-116.

123

Bennett, Larry. (1990). Fragments of Cities: The New American Downtowns and
Neighborhoods. University of Chicago Press.
Bielecka, E. (2005). A dasymetric population density map of Poland. Proceedings of the
International Cartographic Conference. 9–15.
Bivand, Roger S., Pebesma, Edzer J., Gómez-Rubio, Virgilio. (2008) Applied Spatial Data
Analysis in R. Springer. New York, NY.
Blackman, Ivy Q. and David H Picken. (2010). Height and construction cost of residential highrise buildings in Shanghai. Journal of Construction Engineering Management 136. 116981.
Borwn, Paula and Aaron Podolefsky. (1976). Population Density, Agricultural Intensity, Land
Tenure, and Group Size in the New Guinea Highlands. Ethnology 15(3), 211-38.
Brewer, C. A et al. (1997). Mapping mortality: Evaluating color schemes for choropleth maps.
Annals of the Association of American Geographers 87(3), 411–38.
Bruhn, John. (2009). Cohesive Neighborhoods. The Group Effect. Springer Books, New York,
103-25.
Carr, D. B, D. White, and A. M MacEachren. (2005). Conditioned choropleth maps and
hypothesis generation. Annals of the Association of American Geographers 95(1), 32–53.
Casetti, E. (1969). Alternate urban population density models: an analytical comparison of their
validity range. Studies in Regional Science. 105–16.
Chicago Historical Society. (2005). Neighborhood change, 1853-2003. The Electronic
Encyclopedia of Chicago. Retrieved 03/23/2012.
Clark, C. (1951). Urban population densities. Journal of the Royal Statistical Society, Series A.
490–96.
Channick, Herbert S. (2008). Shock City to Chic; the Transformation of Chicago. Journal of the
Illinois State Historical Society 100(4), 383-90.
Coulton, Claudia, Korbin, Jill, Chan, Tsui, and Su, Marilyn. Mapping Residents' Perceptions of
Neighborhood Boundaries: A Methodological Note. American Journal of Community
Psychology 29(2), 371-83.
Daniel, Caroline and Jeremy Grant (2005). Classical city soars above Capone clichés. The
Financial Times (09/10/2005). The Financial Times Ltd. Retrieved 01/17/2012.
Dent, Borden D. (1985). Cartography, Thematic Map Design. Wm. C. Brown Publishers,
Dubuque, Iowa.

124

Duaney A, Plater-Zyberk E. (1993). The neighborhood, the district and the corridor. The New
Urbanism: Toward Architecture of Community. McGraw-Hill, New York.
Eicher, C. L, and C. A Brewer. (2001). Dasymetric mapping and areal interpolation:
implementation and evaluation. Cartography and Geographic Information Science 28(2),
125–38.
Emporis Standards. Data Standards: high-rise building. Frankfurt, Germany. Retrieved
01/17/2012.
Fountain, John W. (2001). “Eyes of Chicago turn to plan for a new landmark.” The New York
Times. (07/28/2001) The New York Times Company. Retrieved 01/17/2012.
García-Ramos, Gisela and Mark Kirkpatrick. (1997). “Genetic Models of Adaptation and Gene
Flow in Peripheral Populations.” Evolution 51(1), 21-28.
Goldsberry, Kirk. (2010). American is not that segregated. Letter to the Editor, New York Times.
www.msu.edu/~kg/nytimes_dotdensity. Retrieved 3/15/2011.
Goodchild, Michael (1986). Spatial Autocorrelation. Concepts and Techniques in Modern
Geography. Geo Books, Norwich, U.K.
Goodchild, M. F, L. Anselin, and U. Deichmann. (1993). A framework for the areal interpolation
of socioeconomic data. Environment and Planning 25(3). 383–97.
Grant, Jill. (2002). Mixed Use in Theory and Practice: Canadian Experience with Implementing
a Planning Principle. Journal of the American Planning Association 68(1), 71-84.
Griffith, Daniel A. (1981). Modeling urban population density in a multi-centered city. Journal
of Urban Economics 9(3), 298-310.
Handley, John. (2009). Looking south- construction in the South Loop is booming, but buyers
are taking their time. Chicago Tribune 07/09/2009. Retrieved 03/23/2012.
Harris, Chauncey D and Edward L Ullman. (1945). The Nature of Cities. The Annals of the
American Academy of Political and Social Science 242, 7-17.
Herold, M., X. Liu, and K. C Clarke. (2003). Spatial metrics and image texture for mapping
urban land use. Photogrammetric Engineering and Remote Sensing 69(9), 991–1002.
Hipp, J. R. (2007). Block, tract, and levels of aggregation: Neighborhood structure and crime and
disorder as a case in point. American Sociological Review 72, 659-80.
Holloway, S. R., J. Schumacher, and R. L. Redmond. (1999). People and place: Dasymetric
mapping using Arc/Info. GIS solutions in natural resource management: balancing the
technical-political equation. OnWord Press, Santa Fe, NM, 283–91.

125

Holt, J. B, C. P. Lo, and T. W Hodler. (2004) Dasymetric estimation of population density and
areal interpolation of census data. Cartography and Geographic Information Science
31(2), 103–21.
Hoyt, Homer. (1939). The structure and growth of residential neighborhoods in American cities.
Washington, DC: Federal Housing Administration. 1939.
Jelinski, D. E, and J. Wu. (1996). The modifiable areal unit problem and implications for
landscape ecology. Landscape ecology 11(3), 129–40.
Kaufman, Mervyn D. (1969). Father of Skyscrapers: A Biography of Louis Sullivan. Little,
Brown and Company. Boston.
Keating, Ann Durkin (1998). Building Chicago: Suburban Developers & the Creation of a
Divided Metropolis. The University of Ohio State Press.
Kern, Leslie. (2007). Reshaping the boundaries of public and private life: Gender, Condominium
development, and the neoliberalization of urban living. Urban Geography 32, 334-59.
Kramer, Gerald. (1983). The ecological fallacy revisited: Aggregate- versus individual-level
findings on economics and elections, and sociotropic voting. The American Political
Science Review 77(1), 92-111.
Lamont, Michèle and Virág Molnár. (2002). The Study of Boundaries in the Social Sciences.
Annual Review of Sociology 28, 167-95.
Langford, M. (2003) Refining methods for dasymetric mapping using satellite remote sensing.
Remotely Sensed Cities. CRC Press, New York, 137-56.
Langford, M (2007). Rapid facilitation of dasymetric-based population interpolation by means of
raster pixel maps. Computers, Environment and Urban Systems 31(1), 19–32.
Langford, M., D. J Maguire, and D. J Unwin. The areal interpolation problem: estimating
population using remote sensing in a GIS framework. Handling Geographical
Information: Methodology and Potential Applications, 55–77.
Langford, M., and D. J Unwin. (1994). Generating and mapping population density surfaces
within a geographical information system. The Cartographic Journal 31(1), 21–26.
Lavin, M. R, and M. D White. (1996) Understanding the census: A guide for marketers,
planners, grant writers, and other data users. Epoch Books, Washington D.C.
Lees, Loretta. (2008) Gentrification and social mixing: towards an inclusive urban renaissance?
Urban Studies 45(12). 2449-70.
LeGates, Richard T. and Frederic Stout. (1999) The City Reader: Third Edition. Routledge
Urban Reader Series. London.

126

Ley, David. (2003). Artist, aestheticisation and the field of gentrification. Urban Studies 40(12).
2527-44.
Liu, X., K. Clarke, and M. Herold. (2006). Population density and image texture: a comparison
study.” Photogrammetric Engineering and Remote Sensing 72(2). 187-96.
Liu, X. H. (2003). Estimation of the spatial distribution of urban population using high spatial
resolution satellite imagery ProQuest Dissertations And Theses 64(8), 3023-198.
Liu, X. H., P. C. Kyriakidis, and M. F. Goodchild. (2008). Population-density estimation using
regression and area-to-point residual kriging. International Journal of Geographical
Information Science 22(4), 431–47.
Lochner, Kimberly A, Ichiro Kawachi, Robert T Brennan, Stephen L Buka. (2003). Social
capital and neighborhood mortality rates in Chicago. Social Science & Medicine, 56(8),
1797-805.
Logan, J., Spielman, S., Xu, H., and Klein, P. N. (2011). Identifying and bounding ethnic
neighborhoods. Urban Geography 32, 334-59.
Maantay, J. A, A. R Maroko, and C. Herrmann. (2007). Mapping population distribution in the
urban environment: the cadastral-based expert dasymetric system (CEDS). Cartography
and Geographic Information Science 34(2), 77–102.
MacEachren, A. M. (1992). “Visualizing uncertain information.” Cartographic Perspective
13(3), 10–19.
MacEachren, A. M et al. (2005). Visualizing Geospatial Information Uncertainty: What We
Know and What We Need to Know. Cartography and Geographic Information Science
32(3), 139–61.
MacEachren, Alan M. (2004). How maps work: representation, visualization, and design.
Guilford Press, New York.
MacEachren, Alan M. and D.W. Dibiase (1991). Automated Maps of Aggregate Data:
Conceptual and Practical Problems. Cartography and Geographic Information Systems
18(4) 221–29.
Maceachren, Alan. (1979). The Evolution of Thematic Cartography. The Canadian
Cartographer 16(1). 17-33.
Macionis, John J. & Vincent Parrillo. (2007). Cities and Urban Life: Fourth Edition. Peasron
Education, Inc. Upper Saddle Ridge, New Jersey.
Mennis, J. (2009). Dasymetric mapping for estimating population in small areas. Geography
Compass 3(2), 727–45.

127

Mennis, J. (2003). Generating surface models of population using dasymetric mapping. The
Professional Geographer 55(1), 31–42.
Mennis, J., and T. Hultgren. (2006). Intelligent dasymetric mapping and its application to areal
interpolation. Cartography and Geographic Information Science 33(3), 179–94.
Monmonier, Mark (1990). Strategies for the Visualization of Geographic Time-Series Data.
Cartographica 27(1), 30-45.
Newling, Bruce. (1969). The Spatial Variations of Urban Population Density. Geographical
Review 59(2), 242-52.
Nobles, M. (2000). Shades of Citizenship. Stanford University Press.
Nyden P, Lukehart J, Maly MT, Peterman W. (1998). Racially and ethnically diverse urban
neighborhoods. In Cityscape 4(2) 261-69. Washington, DC: US Dep. Housing & Urban
Development.
Openshaw S. (1984). The Modifiable Areal Unit Problem. GeoBooks, Norwich, England.
Podagrosi, Angelo et al. (2011). The diversity of gentrification in Houston's urban renaissance:
from cleansing the urban poor to supergentrification. Environment and Planning A 43.
1910-29.
Rankin, Bill. (2010). Cartography and the Reality of Boundaries. Perspecta: The Yale
Architectural Journal 42, 42-45.
Rankin, Bill. www.radicalcartography.net.
Slocum, T. A. et al. (2009). Thematic Cartography and Geovisualization. Pearson Prentice Hall,
New York.
Smith, Neil (1996). The New Urban Frontier: Gentrification and the Revanchist City. Routledge.
New York.
Spielman Seth E and Eun-hye Yoo. (2009). The spatial dimensions of neighborhood effects.
Social Science Medicine 68(6). 1098-105.
Spreiregen, Paul D. (1965). Urban Design: The Architecture of Towns and Cities. McGraw-Hill.
New York.
Steele, Jeffery. (2011). Printers Row writes a new chapter: urban enclaves afford residents
walkable, livable lifestyle. Chicago Tribune: Community Profile. 2/11/2011. Retrieved
3/20/2012.
Taylor, Nigel. (1998). Urban Planning Theory since 1945. Sage Publications. London.
Trewartha, Glenn T. (1953) A Case for Population Geography. Annals of the Association of
American Geographers 43(2), 71-97.
128

United Nations World Urbanization Prospects: The 2009 Revision, Highlights. Hania Zlotnik,
Director, Population Division. New York.
United States Census Bureau. (2010). Census Bureau Geography. Retrieved 10/14/2010.
United States Census Bureau. (2010). 2010 Census Bureau Population and Redistribution.
Retrieved 10/14/2010.
Vojnovic, Igor. (2000). Shaping metropolitan Toronto: a study of linear infrastructure subsides,
1954-66. Environment and Planning B: Planning and Design 27. 197-230.
Waddell, P (2002). UrbanSim: Modeling urban development for land use, transportation, and
environmental planning. Journal of the American Planning Association 68(3), 297-314.
Waddell, P et al. (2003). Microsimulation of urban development and location choices: design and
implementation of UrbanSim. Networks and Spatial Economics 3(1), 43–67.
Weir, H. K et al. (2003). “Annual report to the nation on the status of cancer, 1975–2000,
featuring the uses of surveillance data for cancer prevention and control.” Journal of the
National Cancer Institute 95(17) (2003): 1276-99.
Wen, Ming, Christopher R Browning, Kathleen A Cagney. (2003). Poverty, affluence, and
income inequality: neighborhood economic structure and its implications for health.
Social Science & Medicine 57(5), 843-60.
Wright, J. K. (1936). A method of mapping densities of population: With Cape Cod as an
example. Geographical Review 26(1), 103–110.
Yuan, Y., R. M Smith, and W. F Limp. (1977). Remodeling census population with spatial
information from Landsat TM imagery. Computers, Environment and Urban Systems
21(3), 245–58.
Zillow- Real Estate, homes for sale, recent sales, apartment rentals. www.zillow.com. Retrieved
3/25/2012.

129