University of Southern California Dissertations and Theses

Automating “Ethington Transections”: a new visualization tool

(USC Thesis Other)

Automating “Ethington Transections”: a new visualization tool

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content AUTOMATING “ETHINGTON TRANSECTIONS”:
A NEW VISUALIZATION TOOL
by
Anne Jeanene (AJ) Bengoa
A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOGRAPHIC INFORMATION SCIENCE AND TECHNOLOGY)
August 2016
Copyright 2016 Anne Jeanene (AJ) Bengoa
ii

DEDICATION
I dedicate this document to all my family for their constant support; my loving husband, 2
amazing kids, my encouraging parents and my ever helpful aunts.

iii

ACKNOWLEDGMENTS
I will be forever grateful; to my amazing professor and friend—Dr. Swift, for Dr. Ethington’s
remarkable ideas and continuous support, and to Dr. Vos for his valuable input to this project
both at its launch and terminus. Thank you also to my family and friends, without whom I could
not have made it this far.

i

TABLE OF CONTENTS
DEDICATION ii
ACKNOWLEDGMENTS iii
LIST OF TABLES iv
LIST OF FIGURES v
LIST OF ABBREVIATIONS vii
ABSTRACT viii
CHAPTER 1: INTRODUCTION 1
1.1 Topic Definition 4
1.1.1 Visualizing Historical Data of Place 4
1.1.2 What are Transections? 7
1.2 Motivation 12
1.3 General Objective of Proposed Application 15
CHAPTER 2: LITERATURE REVIEW 17
2.1 A Survey of Linear Visualization and Analysis References 17
2.1.1 Transects vs. Transections 22
2.2 GIS Tool development efforts in the Literature 24
2.3 Previous HGIS efforts 26
2.4 Available tools in on-line repositories 27
2.5 Novelty of proposed project 27

ii

CHAPTER 3: METHODOLOGY 29
3.1 Programming Objectives 29
3.2 Technology: Software Choices and Programming Considerations 30
3.3 Application Development 31
3.3.1 Task Goals Detailed 33
3.4 The Ethington Transection Toolbox Models 34
3.4.1. Transection File SetUp 2.1 34
3.4.2. Visualize Transection 1.4 35
3.4.3. FlipTransection 1.2 37
3.4.4. GeoprocessingTasks 3.0 Test 6.5 38
3.4.5. Organize Polygon Data and Export Table 40
3.4.6. Visualize Polygons 41
3.4.7. Transection Graph_Horizontal (or _Vertical) 42
3.5 Documentation 43
3.6 ETT Evaluation 44
3.7 ETT Efficacy Test 46
3.7.1 Old vs. New: Comparison of Pico-Whittier Transections 46
3.7.2 Washington DC Orange Metrorail 47
3.7.3 Environmental Sensitivity Index along the Intercoastal Waterway 47
3.8 Application Development Issues Encountered 48

iii

CHAPTER 4: RESULTS 52
4.1 Running the ETT 52
4.2 ETT Efficacy Test Results 58
4.2.1 Old vs. New: Comparison of Pico-Whittier Transections 58
4.2.2 Washington DC Orange Metro Line 61
4.2.3 Environmental Sensitivity Index along the Intercoastal Waterway 63
CHAPTER 5: CONCLUSIONS AND DISCUSSION 66
5.1 Considerations for Using the ETT 66
5.1.1 How the MAUP applies to Ethington Transections 70
5.2 Limitations of Current ETT 71
5.2 Addressing Current Issues in Future Work 74
5.3 Additional Future Work 76
5.4 Conclusion 78
REFERENCES 79
APPENDIX A: CONTEXT SENSITIVE HELP DOCUMENTS FROM MODELS 86
APPENDIX B: USER’S MANUAL FOR THE ETHINGTON TRANSECTION TOOLBOX
(ETT) 100
APPENDIX C: MAP DOCUMENTS PRODUCED DURING EFFICACY TESTS 119
Pico Blvd. Transections for Percent White in 1940 and 1960 119
Pico Blvd. Transections for Percent White in 1980 and 2000 120
DC Orange Line Metro Transection for Percent under 17 and Home Ownership 121
Florida ICW Transections for Counts of Reptiles of Special Concern 122

iv

LIST OF TABLES
Table 1. Comparison of Transects and Transections. 24
Table 2. Qualities of a line feature Class that warrant consideration when using the ETT. 67
Table 3. Qualities of polygons that warrant consideration when using the ETT. 68

v

LIST OF FIGURES
Figure 1. Charles Minard's Napoleon's March to Moscow in the War of 1812. ............................ 4
Figure 2. Example of a small multiple map set (Slocum, et al. 2009). ........................................... 6
Figure 3. “Pico-Whittier Transection” (Ethington forthcoming). .................................................. 9
Figure 4. The numbered models in the custom toolbox for the Ethington Transection Toolbox. 16
Figure 5. Traffic flows on Delhi bus routes (Mishra, Parida and Rangnekar 2010). .................... 18
Figure 6. Wildlife movement on Trans-Canada Highway (Alexander and Waters 2000). .......... 19
Figure 7. Elephant density as determined by dung piles at Mount Kenya (Vanleeuwe 2010). .... 21
Figure 8. Catch Transects for multiple years in 3 locations (Watson and Pauly 2014). ............... 22
Figure 9. Flow chart for the development of the Ethington Transection Toolbox. ...................... 32
Figure 10. ModelBuilder view of Model 1. .................................................................................. 35
Figure 11. ModelBuilder view of Model 2. .................................................................................. 36
Figure 12. ModelBuilder view of Model 3 ................................................................................... 37
Figure 13a. ModelBuilder view of Model 4. ................................................................................ 39
Figure 14b. ModelBuilder view of Model 4. ................................................................................ 40
Figure 15. ModelBuilder view of Model 5. .................................................................................. 41
Figure 16. ModelBuilder view of Model 6. .................................................................................. 42
Figure 17. ModelBuilder view of 7. TransectionGraph_Horizontal ............................................. 43
Figure 18. The three test transections used to evaluate model performance. ............................... 45
Figure 19. Contents of Zipped folder for delivery of the ETT. .................................................... 52
Figure 20. ArcMap after model “1. Transection File Set Up” runs. ............................................. 53
Figure 21. ArcMap after Model 2 runs. ........................................................................................ 54
Figure 22. ArcMap after Model 3 runs. ........................................................................................ 55

vi

Figure 23. Example of how the ArcMap windows prior to running Model 4. ............................. 55
Figure 24. An example of ArcMap after Model 4 runs. ............................................................... 56
Figure 25. An example of ArcMap after Model 5 runs. ............................................................... 56
Figure 26. An example of ArcMap after Model 6 runs. ............................................................... 57
Figure 27. An example of a final layout view for a sample ETT run. .......................................... 57
Figure 28. Pico Blvd. transection shown going in 2 different directions. .................................... 59
Figure 29. Zooming in on Pico Blvd. to reveal split lanes along travel path................................ 59
Figure 30. Pico Blvd. Transection in only one direction. ............................................................. 60
Figure 31. The two alternate polygon selections of Pico Blvd. with a portion in small scale. ..... 60
Figure 32. The ordered stops displayed along with the selected polygons for the DC test. ......... 61
Figure 33. The DC Metro Line and 2000 Census Tract data. ....................................................... 63
Figure 34. The initial line feature class of the ICW. ..................................................................... 64
Figure 35. The reptile feature class of the ESI with the Floridian ICW. ...................................... 65
Figure 36. The Input Series for creating a graph via an exposed parameter in a model 7. ........... 73
Figure 37. Mock-up of possible future ETT distributed as a Python add-in toolbar. ................... 76
Figure 38. Mock-up of screen shots while using future versions of the ETT. .............................. 77

vii

LIST OF ABBREVIATIONS
ESI Environmental Sensitivity Index
ETT Ethington Transection Toolbox
GDB Geodatabase
GIS Geographic Information System
HGIS Historical Geographic Information System
ICW Intercoastal Waterway
MAUP Modifiable Areal Unit Problem
SSI Spatial Sciences Institute
USC University of Southern California

viii

ABSTRACT
The goal of this project is to develop an Ethington Transection Toolbox (ETT) to automate,
increase the efficiency, and improve the efficacy of creating “Ethington Transections.” Ethington
presents these hybrid charts/maps to visualize social change in space and time, along urban
streets in his forthcoming book, Ghost Metropolis: A Global History of Los Angeles since
13,000. A new technique for visualizing the act of moving through the landscape over time,
“Ethington Transections” are defined as a cross-sectional sample of data from polygons to
simulate a single, directional line of transit. The objective of this thesis is to streamline the
creation of transections resulting from the input of common polygon-distributed data, and to
share such a tool so that others may benefit from increased efficiencies. The final result is a
custom toolbox in Esri’s ArcGIS ModelBuilder of seven custom models with contextual help,
written documentation and video walkthroughs. This series of models creates an editable map
and graph layout and an organized geodatabase of intermediate outputs that can be reused for
additional analyses or presentations. The ETT shortens the time to complete an “Ethington
Transection” from 8 hours to slightly less than 1 hour. The previously tedious and time intensive
task of creating transections was automated and made accessible to a wider range of researchers,
facilitating new perspectives and interpretations of data. Therefore this toolbox should enhance
the analytic skills of those looking to study how changes occur through space and time along any
linear sample of data to simulate a transit in polygonal datasets.

1

CHAPTER 1: INTRODUCTION
There are many ways in which human beings struggle to understand their place in the world. By
looking back, we often better understand where we are, and illuminate better paths forward.
Since the beginning of recorded history the idea of place has been important (Bodenhammer
2013). In order to fully understand the historical picture the essential question of “Where?” had
to be answered (Bodenhammer 2013). Over time tools have been developed to help us “see” our
human story in different ways. Maps were developed to convey just such spatial information.
With the growth in computers, and specifically the field of Geographic Information Systems
(GIS), the story teller, or modern researcher, could go beyond the map in the ability to analyze
relationships over time in a place.
Transections are a GIS technique developed by Dr. Philip Ethington (Forthcoming) in
order to visualize the act of moving through the landscape over time. Transections can be defined
as a cross-section sample of data from polygons to simulate a single line of transit (Ethington
Forthcoming). The objective of this thesis was to streamline the creation of Ethington
Transections resulting from the input of basic polygon-distributed data, and to be able to share
such a tool so that others may benefit from the increased efficiencies. This thesis project devised
a method to automate the process, thus hastening time to data visualization. The final result is a
custom toolbox in Esri’s ArcGIS ModelBuilder of seven custom models with contextual help,
written documentation and video walkthroughs called the Ethington Transection Toolbox (ETT).
This series of models creates an editable map and graph layout and an organized geodatabase of
intermediate outputs that can be reused for additional analyses or presentations. The ETT can
enhance the analytical skills of researchers studying how changes in almost any kind of spatial

2

data occur through space and time along any linear sample of data to simulate a transit in
polygonal datasets.
The ETT is delivered with all the requisite templates and geoprocessing steps. The
operation of the ETT does not necessitate the distribution of any particular data as it utilizes only
data inputs from the user. However a sample geodatabase (GDB) is made available which
enables a new user to run through the tutorial explained in the accompanying User’s Manual.
This project also utilized both a previously compiled GDB and publicly available data sources
(U.S. Census, NOAA, DCGISopendata, and FWC-FWRI) in order to test the effectiveness of the
ETT. Automation enables researchers to increase the number of transection visualizations
derived and allow for easier comparisons within and between datasets.
Current manual methods call for the data from all the desired variables and data layers for
visualization to be apportioned to a single geographic areal unit configuration. Then, the areal
units along the line being selected are outputted to a drawing program for cartographic
presentation. Next, the data on a variable of interest from those areal units are output to a
spreadsheet program and charted variables of interest, afterward the charts are also moved to a
drawing program to align the chart data to the geographic information for presentation. Then
manual methods are employed to visually align the data points of the graphs with the geography
of the polygons. This manual method is cumbersome and time consuming.
The goals of this project were to provide a toolset and technique to all those looking for
alternative methods of linear visualization, and to broaden the scope of the Ethington Transection
beyond its present usage in urban spatial history. This project automated the above steps for the
creation of Ethington Transections by developing a custom desktop Esri ArcMap Python toolbox
(Esri 2015). Producing a desktop ArcMap Python toolbox worked well to automate transection

3

creation because they are easily produced, distributed, and integrated in to the ArcGIS
application. These ArcMap Python tools are well supported by Esri and widely used by ArcGIS
users of all skill levels., and therefore makes a great platform for the application of the is project.
Other applications have been created to ease processing challenges within ArcGIS and
other GIS software (Bell 2004; Dilts, Yang and Weisberg 2010; Pratt 2000; Rattan, Campese and
Eden 2012; Reed, Boggs and Mann 2012; Reiser 2014; Silva and Taborada 2013; Spalding 2000;
Wheaton et. al. 2012). Geoprocessing tools are developed primarily to simplify or reduce the
completion time of a GIS-based computational task that can be run multiple times, and then they
can be shared with others so that they too may benefit from the efficiencies of the tool (Allen
2011). The development of this tool allows researchers to utilize the advantages of task-specific,
customized GIS data processing tools.
The previously tedious and time intensive task of creating Ethington Transections is
automated and made more accessible to a wider range of scholars. Topics such as ethnic
diversification, gentrification, and economic shifts can be visualized using transections by
examining demographic variables like population by ethnicity, population by age,
unemployment, buying power, and average household size. Alternatively, in the ecological
research sphere, changes in wildlife land use, invasive species distribution patterns, or oceanic
catch densities could also be evaluated with an Ethington Transection.
The structure of the remainder of thesis document: first describes the context of the
project; then reviews the literature on spatial linear analysis, HGIS automation, and GIS tool
development; then confers in detail the specifics of the ETT and its development; next delivers
the results of using the ETT in multiple scenarios; and lastly discusses considerations,
limitations, and possible future work for the ETT.

4

1.1 Topic Definition
In order to adequately define the topic of this project a brief account of some alternate spatial
data visualization methods is presented herein, along with a closer look into what are Ethington
Transections and how they differ from more traditional approaches. Additionally a look into the
application development solutions chosen for this project will be discussed.
1.1.1 Visualizing Historical Data of Place
The role of geography is to distill spatial information so that ideas and patterns become readily
apparent (Montello 2001). The very act of creating a map requires a distortion of reality, where
only a “selective, incomplete view” can be represented on the scaled down two-dimensional
surface at hand (Monmonier 1996). One role of a geographer is deciding what story needs to be
told and then how best to visualize it. This can become more challenging when trying to relate
information that includes time (Gregory and Ell 2007).

Figure 1. Charles Minard's Napoleon's March to Moscow in the War of 1812.

One of the most highly acclaimed spatial visualizations throughout history is Charles
Minard’s depiction of Napoleon’s March to Moscow during the War of 1812 as reproduced in

5

Figure 1 (Wikimedia Commons 2105). It goes beyond a simple map or graph and endeavors to
deliver large amounts of information allowing for multiple comparisons and true “graphical
excellence.” Edward Tufte (2001, 51) posits, “graphical excellence is that which gives to the
viewer the greatest number of ideas in the shortest time with the least ink in the smallest space.”
To truly deliver an impactful message all modern scholars should struggle to do the same.
The story of a place can be told in many ways in order to synthesize and illuminate
different facets of the complex web of information experienced at that location. Just as altering
the angle, zoom or aperture of a camera readily changes the image it is shooting, so can altering
the tools, data, and perspectives of a story change our “image” of a place. When GIS is
employed to tell this story, “innovative ways to collect, manage, model, analyze, interpret, and
display geographic information” (Offen 2013, 568) are developed. HGIS becomes the tool to
facilitate this shifting perspective by incorporating the essential added dimension of time
(Raymond 2011). However no matter the quantity of visualizations GIS and “big data” tools can
deliver, “there are few that can assemble the data collected by Minard and create a more
sophisticated representation” (Plummer 2014, “How modern visualizations” para 1).
One of the most common ways researchers use GIS outputs to represent pattern
distribution in a place is with a series of choropleth maps that view the panorama from a typical
map scale delivered as a series of small multiples. In the example of this formula seen in Figure
2, climate conditions of California are shown in six small multiples (Slocum, et al. 2009). The
top ones are in a continuous raster format, while the bottom indicates the results distributed by
county. This arrangement allows the viewer to draw conclusions about the relationship between
temperature and precipitation in the state, as well as to illustrate the Modifiable Areal Unit
Problem (MAUP) (Peterson 2011). In the small multiple, three or more maps or images are used

6

in a sequence to display different data from the same geographic space (Tufte 1990). This
composition allows all the maps to be viewed semi-simultaneously in order to draw conclusions
by discerning inter-map similarities and differences (Peterson 2012). This gives the viewer the
impression they are watching the polygons on the ground shifting colors as if viewed from a
stationary point in orbit above the Earth (at for instance 10,000 feet). These visualizations have
proven useful and are able to reveal patterns in the changes of numerous variables (ethnicity,
economic status, land cover, rainfall, etc.) (Slocum, et al. 2009).

Figure 2. Example of a small multiple map set (Slocum, et al. 2009).

7

As useful as this method is, Ethington supposed that it did not adequately reflect the
differing human experiences of place (Forthcoming). A person does not acquire a city’s character
solely from the sum total of statistical facts over time (Lynch 1960). And one certainly does not
receive this data in an instant, as if downloaded directly to our psyche. Also, one does not
typically experience the urban landscape from a bird’s eye view.
Instead, the observer of a city creates a relationship with it and attempts to understand it.
We have to interact with the landscape in a spatially ordered manner. The most common way
that people experience a place is by moving through it (Lynch 1960). This is exemplified when
driving down a street, walking along a river bank, or riding on a train. At the beginning of this
process, only the small piece of the puzzle in which a person stands can be visualized. Over time,
the known puzzle pieces come together to create a distinct picture of a city. Just as someone
working a puzzle can imagine what the form of the missing pieces should be, the city wanderer
can work similar strategies on their mental urban image. Even if we don’t have all the pieces—
have not experienced every part of the city—we know what the pieces should contain to
complete the image. Thus we have the confidence to navigate the “in between spaces” that were
previously blank in our knowledge base. Lynch (1960) famously termed this property of cities
“imageability”.
1.1.2 What are Transections?
In order to visualize the act of moving through the landscape Ethington drew inspiration from the
likes of Minard and Tufte to develop a new spatial data visualization technique—the transection
(Forthcoming). He defined it as “a cross-section of the metropolis on a single line of transit”. By
creating a transection based on popular thoroughfares one can experience both the impression of
a commute and of transits between segregated spaces. This transection technique capitalizes on

8

the advantage of using GIS—the opportunity to develop new analytical techniques to best
represent spatial data (Gregory and Ell 2007).
The first step to creating an Ethington Transections is to select a significant transit
through the region (Ethington Forthcoming). It is desirable to choose arteries or paths that have
permanent relationships with the landscape. Then all the areal units that touch that path are
queried for their data. The data are then organized in the direction of the path. Next a variable
such as % white non-Hispanic, median household value, or the like, is plotted as a standard line
graph. Graphs can be repeatedly generated on different variables for simultaneous display and
cross comparison. The plots are standardized such that the distance between the data points
corresponds to the physical/spatial distance traveled along the path represented in the transection.
An example of this process can be seen in Figure 3, “Pico-Whittier Transection” from “The
Spacetime Transection: Pico-Whittier, Lakewood-Rosemead, and Sepulveda” in Ghost
Metropolis: Los Angeles Since 13,000 BP (Ethington forthcoming).

9

Figure 3. “Pico-Whittier Transection” (Ethington forthcoming).

10

Ethington presents the first charts made along his unique street-derived transections in
Ghost Metropolis: Los Angeles Since 13,000 BP (Forthcoming). These transections plot ethnicity
and economic data for 1940, 1960, 1980 and 2000. Transections were chosen along the east-
west streets following Pico Blvd and Whittier Blvd, the north-south oriented Lakewood Drive
and Rosemead Avenue, and the meandering Sepulveda Blvd. The Pico-Whittier Transection can
be seen in Figure 3 as an example of the technique. Graphs from these transections reveal that
property value patterns tended to be more stable over time than the still relatively stable pattern
representing the percentage of white residents (Ethington Forthcoming).
1.1.4 What is Inscription?
Inscriptions are conceptual imprints left on a landscape from the past political and cultural
activities (Ethington Forthcoming). One example would be how the initial delineation of ranchos
in LA was based on previous indigenous settlements, and whose pattern can still be seen in the
major centers of population today .In an effort to discover and convey these theories, one of
Ethington’s goals is to find and visualize these inscriptions; principally for Los Angeles, but
ultimately for other cities and social spaces as well. Ethington Transections grew out of this
particular objective.
By examining the institutional inscriptions left by the changing Spanish, British, Spanish
(again), and American governmental controls, Baldwin was able to use transections to visualize
the remarkable persistence of land usage in in St. Augustine, Florida, especially as it related to
attributes of the parcel geometry (2014). Transections were used to visualize the parcel
geometry shifts of this area (Baldwin 2014). The Minorcan Quarter showed a persistent style of
cultural, economic and government usage patterns that was difficult to explain at first with other
methods (Baldwin 2014).

11

1.1.3 Finding patterns in the history of Los Angeles
The large and complex network of communities making up Los Angeles is an ever-enticing,
ever-ambivalent, and increasingly multicultural megalopolis (Brook 2013). Los Angeles
represents a diverse quilt of people living within the landscape that have evolved through history
(Ethington Forthcoming). The landscape and sense of place are key. This makes Los Angeles an
interesting and rich source for historical research. As Ethington explains:
What holds it together in the imagination (which is always partly visual) are
mental images of a whole city, imaginary illusions of a whole to cover the actual
reality of a million separate parts. What holds it together functionally is the state,
the institutions of civil society, and the infinite daily practices of its millions of
residents. (Ethington Forthcoming)
Pulido, Barraclough and Cheng assert that Los Angles is made of just more than its
people, but also its ghosts (2012). These ghosts of the past can be felt by exploring the historical
places and moments that still haunt Los Angeles today. Ethington’s forthcoming book Ghost
Metropolis: Los Angeles Since 13,000 BP delves into the history of Los Angeles County with
data rich cartographic visualizations. The themes explored—demographic patterns, institutional
inscriptions, and transections—strive to unearth and visualize the relationships of these ghosts of
the past to the current Los Angeles landscape (Ethington Forthcoming). All of his presented
work showed unique and powerful methods for analyzing Ethington’s primary interest,
“historical change over time and space”—in its current context of Los Angeles County. The
book is an attempt to tell Los Angeles’s history of the “inscription of institutional forms into the
landscape.” Ethington argues that such inscriptions give this area its unique character, and more
precisely, created the spaces that shaped and continues to shape the lives of all Angelenos.

12

1.2 Motivation
It was decided that this particular visualization technique (transection) produces valuable insights
to the study of how changes occur through space and time and thus would be desirous for
multiple uses across many fields of spatial inquiry. However the current method for conducting
the Ethington Transection visualization (detailed in the next section) is long, cumbersome and
prohibitively difficult for use by researchers lacking a professional level knowledge of GIS
software. Therefore the primary motivation for automating a visualization using transections is
to broaden the use of transections by the research community.
1.2.1 How to perform Ethington Transection visualization without the tool
In order to properly understand the scope of the project and to accurately compare the manual
method of creating transections to that of the automation procedure, the step by step instructions
for a manual Ethington Transection visualization are sketched out in this section.
Prior to creating a transection, the initial data must be assembled to standardize the areal
units. The specific data assembly for Los Angeles was described by Ethington in his analyses of
the ethnic shift between 1940 and 2000 (Forthcoming).
First, assemble all the data of interest of a given visualization. There may be situations
when this data occurs in different geographic units in different time periods or data sets and
therefore must be standardized to a specific/chosen set of areal units. Then, all the data across the
variables must be fitted to the chosen geography by area apportionment. Data from the U.S.
Census Bureau acquired at the census tract level is an example, while custom geographic areal
units can be devised to meet research needs. For instance, a study area can be divided based on
school districts, watersheds, or parcels. In creating any level of areal unit, any inconsistencies

13

must be reconciled, such as population occurring in unincorporated areas or multiple polygons in
fragmented zones.
Second, all the variables required for the visualization need to be reported using the same
criteria and codes. Any variations in criteria or coding schemes must be reconciled and well
documented.
After the communities are designated and each contained within a polygon, a transection
can be generated following the steps listed in Ethington (Forthcoming). For example:
1) In a GIS, a line is created that will define the transection along which the areal
units will be sampled.
2) Every census tract (any polygon) adjacent to the transection is selected, sampled,
and the data is exported in a tabular format (ex. Excel spreadsheet).
3) The selected polygons (census tracts), their labels, and the transection line vector
are exported to a graphics format (such as Adobe Illustrator’s vector based
encapsulated post-script [EPS]) to create a map and spatial reference for the
transection’s final graphic.
4) The attribute data tables associated with the selected, adjacent polygons are then
arranged in an orderly E-W or N-S directional order, whichever is most
representative of the transection selection.
5) Data from any selected variable along that sequence is plotted on a line graph. A
separate line graph is created for all desired variables.
6) All the line graphs are imported into an image editor such as Adobe Illustrator so
that the layout and format is adjusted to the scale and position of the map data.

14

7) Appropriate map elements are added to the graphic are made such as a legend,
scale bar, labels, and title.
Despite this standard recipe for conducting an Ethington Transection visualization, there
are still assumptions and choices that must be made while assembling the final product. One such
decision is what to do if polygonal bounded data sits along the line/path in irregular ways such as
overlapping the line, or occurring just above or below it; should all these polygons be included or
only certain ones? Decisions must also be made as to how the polygons and their associated data
shall be ordered in the map and corresponding line graphs.
The information gleaned from the transections paints a picture of Los Angeles that is
difficult to acquire elsewhere. It represents a unique and relevant view of a place that is linear,
sequential and directional. It is a vision worthy of exploring further and implementing in other
metropolises, or even in other realms of inquiry all together.
The current methods of Ethington Transection visualization are time consuming, tedious
and difficult to implement. Although it would be possible for others to create transections using
the current methods, the daunting workload would deter all but the most ardent researcher with
extensive GIS skills. Consequently this project proposes to devise a method to automate,
increase the efficiency, and prove the efficacy of Ethington Transection visualization. By
creating a tool to use within a GIS environment, the time to data visualization is shortened and
the technique is available to be shared amongst other researchers (Allen 2011).
Whereas transections have already been applied to examine the issues of ethnic
segregation, median household values, and parcel geometry (Baldwin 2014, Ethington
Forthcoming), additional specific aspects of urban history such as blight, gentrification,
architectural structure (building height), immigration integration, drug use, crime trends, and loss

15

of green space could also be evaluated. Urban historians who maintain a database of information
about their metropolises of interest could add this tool to their “analysis toolbox”, particularly
when a metropolis maintains a strong connection to distinct corridors that shaped the changes of
their development. For example the Ohio River through Cincinnati, OH which acts as both a
physical and governmental demarcation between the states of Ohio and Kentucky which has led
to drastically different settlement patterns on either side through history, especially as the
influence of the river as a transportation hub waned in the late 20th century. To discover how
this path has led to differential establishment of race, economic classes, and land use along the
river course would be interesting. Another example is the development of South Florida’s “Gold
Coast” which initially followed the Intercoastal Waterway, then later shifted to Flagler’s
influential railway, then shifted again to street based transportation. The influx of various
migrant groups to the area and the different cultural connections that they brought with them
could be visualized against ethnic integration, economic indicators and land use.
It is also foreseen that this technique would appeal to researchers beyond urban
historians. For example, environmental managers could show how the density of trees along a
logging road has changed over time, how ecological classifications have changed along an ever
more utilized highway corridor, how dominant flora change seasonally on a hiking trail or how
fish densities have changed within a river throughout various management practices. Any
researcher that is looking for alternative linear visualization techniques to examine their
geographically distributed data might be able to utilize the ETT.
1.3 General Objective of Proposed Application
The potential users of the tool are researchers that are interested in illuminating a characteristic
(i.e., population, land usage, species count, etc.) along a representative track from their areas of

16

interest. The tool was designed so that minimal GIS and graphic design technique are required
to achieve visually appealing results that allow for effective analytic comparisons. The tool was
initially tested on data from the Los Angeles area to make it comparable to previously performed
transects as described Chapter 4. The automation was accomplished by creating numbered tools,
which run from a custom toolbox in ArcMap that starts from a blank map (Error! Reference
source not found.).

Figure 4. The numbered models in the custom toolbox
for the Ethington Transection Toolbox.

17

CHAPTER 2: LITERATURE REVIEW
In order to put this project in perspective, this chapter reviews the literature about previous
similar GIS programming efforts. Acknowledging the derivations of transections in the HGIS
community and simultaneously looking towards its use in the broader research community for
spatial visualizations, the topics proceed from linear visualization and analysis techniques to
basic tool development, while briefly discussing previous HGIS efforts and a review of GIS tools
previously made available to the public for use by the GIS community. To close chapter, the
distinctions that make this project unique are discussed.
2.1 A Survey of Linear Visualization and Analysis References
Research into the literature revealed that generally linear analysis techniques create an equation
model of line pattern or graphs data, often employing statistics. If any visual spatial reference is
given at all it is a map juxtaposed with the graph. It is then left for the reader to make the
connection between the spatial relationship and that of the distance on the graph. Articles that
exemplify slight variations on this strategy in different fields are described next.
Mishra, Parida and Rangnekar quantify and analyze the traffic noise emissions along a
bus rapid transit corridor in Delhi (2010). Field measurements were carried out to understand and
assess various aspects of the impact of the bus rapid transit system corridor on land use and the
social lives of residents and road users. The presentation of some of this data is seen in Figure 5
where the horizontal line graph is juxtaposed with a map in a characteristic arrangement. The
remainder of the data presentations—of noise level, indicators of resident health, and additional
traffic flow—are line or bar graphs. This particular dataset would be well-suited to visualize in
an Ethington Transection as it of a transit path through varying sectors of the urban landscape.

18

Figure 5. Traffic flows on Delhi bus routes (Mishra, Parida and Rangnekar 2010).
Investigating a different set of parameters associated with well-traveled roadway,
Alexander and Waters monitored wildlife movement across and adjacent to the Trans-Canada
Highway (2000). The annual average daily traffic rate of the Trans-Canada Highway is 14,000,
and the adjacent Highway 1A is 3000 as they traverse Banf National Park, in Alberta, Canada.
Animal tracks were observed crossing roadways and on transects adjacent to roads for wolves,

19

cougar, lynx, wolverine, marten, elk, deer, sheep, hare, and red squirrel relative to road types,
and it was determined that the highway was indeed a barrier to movement for all species. The
figure most connected to the spirit and purpose of an Ethington Transection can be seen in
Error! Reference source not found., where the map is labeled and two small horizontal line
graphs of wildlife movement are inset. The remainder of the article depends on tables of data
and summary statistics to make its case. Again this seems a perfect candidate for an Ethington
Transection whereas this research actually used a well-established transit path as its locational
determinant for data collection, only failing to make the leap to visually connect the data to the
spatial distribution.

Figure 6. Wildlife movement on Trans-Canada Highway (Alexander and Waters 2000).

20

Some other studies investigated how to model water infiltration through bedrock and
identify steep concentration gradients (Maier, Flegr, Rügner, and Grathwohl 2013), and the
yearly seasonal depth distribution patterns of an invasive macroalga via a transect through a
typical Sargassum bed in Limfjorden (Thomsen, Wernberg, Stæhr, and Pedersen 2006). This
second example contains one graph series showing variables such as species richness, %
coverage, and % substrate mapped against distance from shore. At least this representation is
scaled to match the landscape, though again the final step of connecting the spatial visually to the
data is missed.
A particularly entertaining study mapped seasonal distributions of elephant dung with
transects (Vanleeuwe 2010). A selected number of line-transect were used to collect data on
elephant dung piles that were then analyzed using the generalized linear model. The results of the
analysis was used to build explanatory models and distribution maps (heat maps) which became
powerful tools for patrol planning and land-use management by locating the areas of high
elephant density and the habitats they move between. Here the data collected on a line was used
to interpolate results to the larger geographic area of Mount Kenya National Reserve that
encircles Mount Kenya National Park. Because of the rugged terrain and inaccessibility it was
near impossible to distribute transects evenly over the entire habitat as would typically be done,
instead walkable transects were created that were as similar in nature as possible in length and
relationship to natural features (such as rivers), but that managed to traverse the diversity of the
possible habitats in the reserve. Consequently these transects take on more of the feel of
transection. However as the purpose of a transection is to display polygonally distributed data in
a linear format, the purpose of this transect was the inverse—to interpolate the linearly collected

21

data to a larger geographic area that was visualized as heat map distributions for the entire
Reserve (Figure 7).

Figure 7. Elephant density as determined by dung piles at Mount Kenya (Vanleeuwe 2010).
One final mention of a study using linear techniques. Just as Ethington created
transections to visualize the patterns he wanted to explore in urban historical data, Watson and
Pauly created ‘catch transects,’ a novel intuitive approach for the representation of fisheries
catches within profiles perpendicular to the coast (2014). These ‘catch transects’ show where
catch is extracted in the water column on plots of bathymetry versus distance offshore and thus
allow for spatial representation of the catch density of pelagic and benthic fisheries in heat maps
(Figure 8). Hence, they allow direct visual comparison of the intensity of fishing through time
and space, very much in the spirit of an Ethington Transection.

22

Figure 8. Catch Transects for multiple years in 3 locations (Watson and Pauly 2014).
2.1.1 Transects vs. Transections
The search term “transect” yielded 22,694 results just from the general search engine employed
by the USC library system, and thus it would be impossible to assert a thorough understanding of
them all. Therefore with the awareness that no search can ever be claimed to be thoroughly
exhaustive, and after sampling several hundred articles on the subject of transects and spatial
linear analysis and visualizations, it will be stated here that no exact replicate of the Ethington
Transection technique has been found reported in the literature. Therefore it will be instructive
to discuss the differences between and Ethington Transection and traditional transects a summary
of which can be evaluated in Table 1.

23

By its very nature, an Ethington Transection is linear, sequential and directional. Most
uses of transects only call on their linear quality and fail to incorporate any sequential or
directional information, particularly in their data display. Transections make use of a pre-existing
human path of travel over the landscape that has purpose and meaning, and very likely
repeatable, versus a random or systematic line that is conjured solely for the use in the study.
Other data displays do not typically align with cartography even though admirable examples of
this concept have been historically available such as the previously mentioned Minard’s
visualization of Napoleon’s Russian campaign (Figure 1; 1869). Instead the map and graph are
simply set beside each other with the left to reader to ponder the spatial relationship, unlike
transections which make the effort to graph data that is spatially aligned and scaled to match
map.
The purpose of Ethington Transections are to visualize how data density drawn from a
polygonal distribution changes as you move along a previously defined path through the
landscape versus that of transects which are to define probable density distribution of an area by
sampling a “representative” linear path through that landscape. These hint at the differing
underlying assumptions of the two techniques; transects use point data collected along a line to
interpolate to a broader geographic area, whereas transections visualize polygon distributed data
along a specific path of transit.

24

Table 1. Comparison of Transects and Transections.
Transects Transections
Uses point data collected along a line to interpolate
to a broader geographic area such a polygon
Uses polygon distributed data to visualize a
specific linear path of transit
Novel path; A line randomly or systematically
generated in the landscape that is irrelevant to
repeated measures and generally only defined and
meaningful within the bounds of the study
Pre-existing path; A path of travel through the
landscape that has purpose and meaning and
likely repeatable (i.e. streets, rivers, railroads)
Cartography separate: Graph and map are set
proximate to each other without regard to scale
and visual alignment
Integrated with cartography: Graphed data is
spatially coordinated, aligned and scaled to match
map
Purpose is to define probable density distribution of
an area by sampling a “representative” linear path
through that landscape
Purpose is to visualize how data density drawn
from a polygonal distribution changes as you move
along a previously defined path through the
landscape

2.2 GIS Tool development efforts in the Literature
Other similar computing tool development projects, outside of linear analysis and visualization,
have also been reported in the literature. A shorter term (30 year) time series of the sea ice
benefited from the development of a Java-based tool to ease access and visualization of the
National Ice Center’s (NIC) abundant data (Tang and Wong 2005). These tools were developed
to facilitate the analysis and statistical visualization of the NIC dataset via the web. An HGIS
tool for urban planning decisions, iCity, utilized irregular cellular automata modeling and was a
user-friendly tool designed to be embedded for use within a GIS desktop application (Stevens,
Dragicevic, and Rothley 2007). CHaMP is a similarly webpage embedded tool to transform
environmental monitoring survey data that was created as an ArcGIS Add-In with a user-friendly
interactive interface (Wheaton et. al. 2012). The Florida Panther Habitat Estimator tool
automates the geo-processing tasks used to apply a Mahalanobis distance (D 2) statistical habitat
model to quantify the effects of land-use changes on panther habitat, and was designed for use by

25

persons with only modest GIS familiarity (Murrow et. al. 2013). The System for the Prediction
of Acoustic Detectability GIS tool (SPreAD-GIS) is implemented as a toolbox in ArcGIS that
was written in Python in order to model noise propagation in natural habitats by using commonly
available datasets on land cover, topography, and weather conditions (Reed, Boggs and Mann
2012). The Beach Morphodynamic Model tool (BeachMM) integrated two prediction models for
coastal processes— Simulating Waves Nearshore (SWAN; SWAN 2015) and XBeach
(Roelnivik et al. 2009; Roelnivik 2010)—into one Python script for use within ArcGIS (Silva
and Taborada 2013). BeachMM was shown to greatly simplify dataflow effort, reduce human
error, and provide a dynamic visualization of these coastal prediction models.
A highly ambitious project developed a set of geoprocessing tools for integration of
several platforms to aid in marine geospatial analyses. From within ArcGIS ecologists can utilize
the Marine Geospatial Ecology Tools (MGET) that provide user-friendly access to the advanced
analytic methods available in ArcGIS, Python, R, MATLAB, and C++ which are extensible,
powerful, easy-to-use, and open-source (Roberts et. al 2010). The aim of these tools was to put
the techniques in the hands of the average marine ecologist so that they did not have to perform
the programming or learn every software package to accomplish the analyses.
A very similar approach to the one taken for this project can be seen implemented for the
analysis of walkability in Canada (Rattan, Campese and Eden 2012). The complete walkability
analysis workflow can easily be repeated due to automation provided by ArcGIS ModelBuilder
which was delivered in a custom toolbox. As data are updated, only simple changes in input
parameters are required to run through the four models and perform new analyses. The parallels
to HGIS continue with its use of ArcMap’s geocoding tools and the Esri Network Analyst

26

extension (Esri 2015a). The automation of the first four steps of the workflow reduced the
necessary time, effort and technical GIS skill required to conduct the analysis.
Additionally ArcUser magazine, over the last 15 years, has featured many uses of
ModelBuilder to automate workflows varying from transportation networks, fire hazards, flood
heights (Bell 2004; Dilts, Yang and Weisberg 2010; Pratt 2000; Spalding 2000). All these
articles showcase how ArcGIS ModelBuilder can streamline the analysis process to reduce time,
effort and skill level required, which is the goal of this particular thesis project.
2.3 Previous HGIS efforts
Technical work in history research in the GIS field has involved the creation of projects that
digitize historical documents and create maps to allow for spatial analysis. Just such an example
is in the HGIS of the Southern Great Plains created to digitize the erosion maps from 1936-1937
(Cunfer 2011). Another example is where the urban historians gathered data on the single
tradesmen of Victoria, BC’s building boom circa 1891. This case study combined qualitative
and quantitative data on 2,000 urban working men to create a spatial database as a research tool
for labor historians (Dunae et. al. 2013). And yet another HGIS was created based on the
European railways during 1830-2010 in order to analyze the relationship between population
distribution and railway development (Morillas-Torné 2012). A case study in Seattle, WA on the
effects of the removal of the 245-foot Denny Hill in 1906 and 1935 developed an HGIS by
digitizing the study site evidence between 1893 and 2008 (Raymond 2011).
In addition to the development of HGIS, there has been work in tools developed to
facilitate HGIS analyses. Audisio, Nigrelli, and Lollino (2009) created a new HGIS tool
implementing a systematic strategy for analyzing historic geologic instability processes that
utilizes the framework of a tool fashioned for swift and methodical ingestion of records into a

27

GIS. This tool adds the ability to ingest visual information (such as maps and aerial photos), and
allows the analysis of instability phenomena to include the effects of intense rainfall events. The
new information gathered with their HGIS tool allowed the researchers to conclude that the
amount of catastrophic events was increasing over time and thus inform policy makers so that
they could properly prepare for future situations.
2.4 Available tools in on-line repositories
A slightly different execution can be seen in some of the apps available in the ArcGIS Online
Marketplace. In particular, Esri Insights operates ways that parallel this proposed tool. A user
can generate info-graphics of demographic data by placing a pin (with buffer), drawing a line
(with buffer), or delimiting an area using a simple configurable and interactive web-based tool
(Esri 2014).
At the time of this writing, a search of GitHub’s ArcGIS repositories yields 1,626 results,
296 of which are Python based (GitHub, Inc. 2015b). Tools (and toolboxes) can be found that
range from making it easier to work with NASA remote sensing data, to acquiring Tweets within
a radius of individuals features, to just making it easier to batch move files in ArcCatalog. Of
note is the ArcGISCensusDownload Python toolbox which automatically downloads relevant
Census data based on a user’s specified area of interest (Reiser 2014). This particular tool would
be helpful preparing data for areal reapportionment prior to running the Ethington Transection
Toolbox.
2.5 Novelty of proposed project
The proposed project is similar to previous linear analysis and visualization efforts in that data
for analysis are acquired along a linear path. Although this one differs in that the line is drawn

28

along a transit path versus other random or systematic style transects, and aligning the data
points to locations on a coordinating map further enhances the technique’s visualization.
Additionally the use of ArcGIS ModelBuilder to automate workflows has been well documented
in the scientific literature and trade journals as a method to increase efficiency and broaden the
audience of available users for particular analysis and visualization techniques. This project is
the first attempt at a user-friendly, interactive automation tool for linear visualizations
specifically for creating Ethington Transections at the time this project was conducted. The
creation of this toolbox is truly a synthesis of techniques and theory that has not been achieved
before.

29

CHAPTER 3: METHODOLOGY
The spatial analysis and mapping technology currently available are continuously evolving. In
general, a spatial analysis which consisted of hundreds of steps and took days to complete, from
theory to process to final graphics, is now automated for ease of use. In this study, the primary
goal of this thesis project was to employ available GIS technology to simplify and automate an
existing lengthy set of procedures originally being executed in Adobe Illustrator, Microsoft
Excel and Esri ArcGIS. The Ethington Transection (Ethington Forthcoming) has been developed
from a basic theory of visualization for “how a human travels through space” approach of a
geographically distributed variable into a compact digital spatial visualization and mapping
methodology which is described in this chapter. Since the foundational steps of the process were
already being performed using Esri, a popular GIS industry standard, and since ArcGIS has
additional powerful automation tools built-in, this was deemed the perfect technology to
implement this thesis project.
3.1 Programming Objectives
The primary programming objective for this project was to design a desktop application that can
be executed from within Esri ArcGIS that dramatically simplifies and shortens the time required
to perform an original Ethington Transection (Ethington Forthcoming). The objective is to allow
users to easily select a path through a geographic area and to create an Ethington Transection
graphic (map and line graphs in a standard layout) and decrease the time it takes to perform these
tasks. To meet this objective, the intent was to code and implement this desktop application with
a user-friendly interface that does not necessitate additional software requirements of the user
beyond Esri’s ArcMap.

30

An important motivation for this effort is to put the Ethington Transection visualization
technique into the hands of more users, so that others can better understand the variable within
the spatial context as they change across a landscape. Ensuring that the Ethington Transection
Toolbox is user-friendly and timesaving, increases its appeal to a broader audience. Removing
requirements to install and license additional software other than Esri to perform the
visualization should also reduce barriers to its use.
Thus for the purposes of this study, a custom toolbox within ArcGIS 10.3 was created
using ArcGIS ModelBuilder to allow for easy sharing of application within the ArcGIS user and
developer communities. The tool also requires ArcGIS Network Analyst extension version 10.0
or higher for sequencing geoprocessing tasks along a user-selected spatial path or transect. The
Ethington Transection visualization variables were exposed so that each tool user can easily
introduce unique input data for which transections could be created. This means that the tool can
be used anywhere (where input data is available), thus is independent of location.
3.2 Technology: Software Choices and Programming Considerations
Esri ModelBuilder allows the developer to connect a long string of geoprocessing tasks into a
single tool, or step for the tool user. By laying out blocks resembling a flowchart within the
ModelBuilder interface, a developer can visually connect the tasks that represent the underlying
code (Allen 2011). The ETT was divided into seven ArcMap models to allow for user
modifications between steps, and multiple runs of the same data while changing selected
variables (see Figure 4).
Historically, Ethington Transections were created using Adobe Illustrator for creating
map layouts and Microsoft Excel for generating graphs of results. Illustrator is similar to
ArcMap only in regards to map creation or layout functionality. Conveniently, the ETT

31

developed for this thesis allows users to export results to file formats suitable for graphics and
presentation programs such as Illustrator and Microsoft PowerPoint. The graphing previously
done using Microsoft Excel is now accomplished through ArcMap graphing tools, where again
users also have the option of exporting the results to other charting or graphics software for
further refinement.
3.3 Application Development
Figure 9 provides an overview of the Ethington Transection Toolbox application development
process. The general tasks are listed down the center in order performed. The itemized goals of
each task are detailed immediately to the right of each task. The Application Development
functionality used to accomplish the enumerated tasks are listed on the left of each task in a font
color that coordinates with the task numbers. The creation of the ETT Evaluation (4),
Documentation (5), Zip ETT for Delivery (6) and ETT Efficacy Test (7) are listed as the last
three steps of the process in the diagram, however behind-the-scenes these processes
continuously occur until the completion of the project.
The tasks in Figure 9 are ordered so that the majority of the tool’s creation progress is
completed early on in the application development process. This allowed for the maximum
amount of time to be devoted to testing and validation of the results of running the tool. In
hindsight this was a good move, considering the challenges in coding the beginning and ending
models as a series of geoprocessing tasks.

32

Figure 9. Flow chart for the development of the Ethington Transection Toolbox.

33

3.3.1 Task Goals Detailed
The individual goals are detailed next in order to better understand the specifics of each task in
the Application Development Flow Chart (Figure 9). For the first task, “Automate Geo-
processing Tasks,” the goals were to: establish a geodatabase workspace; select the transection
line, measure its length, and export as feature class; select adjacent areal units, export them as a
feature class, assign an order to polygons, and establish geographic centers along transect line;
and select the columns (variables) of interest and extract all relevant data to a feature class and
exported table. The goals of the second task, “Automate Transection Visualization,” were to:
establish a virtual transection line as a straight line space for graphing; assign areal unit centers
along a virtual line, and get measurements of position along that line; and plot the desired data.
“Automate Display of Transection” was the third task whose goals were to: display data plots
with labels and predefined symbology; add transection and areal unit selections map with labels
and predefined symbology; and add a title and brief explanation. For the fourth task, “ETT
Evaluation,” the goals were to have the: developer test the entire workflow of ETT with different
data scenarios; developer refine the ETT as issues arise from testing; developer’s committee
members test the entire workflow of the ETT; and developer refine the ETT based on comments
from committee. The fifth task’s goals, “Documentation,” were to: label major steps within each
model diagram; construct context sensitive help (Appendix A); create a “User’s Manual” with
step by step instructions (Appendix B); and produce short instructional “how to” videos (links
available in Appendix B). The goals of the sixth task, “Zip ETT for Delivery,” were to: collect
all seven modules in one custom toolbox; create a Geoprocessing Package within ArcMap of all
necessary supporting files; and create a compressed folder (zipped) of ETT file for cloud/ web-
based delivery (Figure 19). Lastly the two goals for the seventh task, “ETT Efficacy Test,” were

34

to test the ETT on current Los Angeles data, and then to test it on two other data types from
public data sources. One was visualizing Washington DC’s population characteristics along the
Orange Metro Line, while the other was plotting the species diversity of the Environmental
Sensitivity Index along the FL Intercoastal Waterway.
3.4 The Ethington Transection Toolbox Models
The seven tools within the ETT are detailed next. Each numbered tool narrative contains a
description of its function, the necessary user inputs and a listing of its outputs. The numbers
following the tool’s name indicate version numbers of that particular component. These
designations were necessary during application development to maintain organization while
testing multiple programming strategies. All actions are executed in ArcMap unless otherwise
indicated in the narrative. All seven help documents detailing parameter input criteria can be
found in Appendix A.
3.4.1. Transection File SetUp 2.1
This model will help collect all the files and create the beginning file structures necessary to
create the Ethington Transection illustrations (Figure 10). This model will run from the catalog
window starting with a new blank map open in the map window. Each time this model runs a
unique set of names for the folder must be entered, otherwise errors will be generated
proclaiming that the file already exists. In order to run this tool a second time with same names,
the user must delete the files created during the first iteration of this tool. The two inputs for this
model are; a polygon file (shapefile or feature class) of the areal units containing data on the
variable desirous of analysis, and a file (shapefile or feature class) containing the line feature to
be used as a transection. The outputs generated are a new file geodatabase with one dataset and
two feature classes. The feature classes are copies of the two inputs. The two feature classes are

35

placed as layer files in the previously blank map and projected according to the coordinate
system designated in one of the parameters.

Figure 10. ModelBuilder view of Model 1.

3.4.2. Visualize Transection 1.4
This model is intended to visualize the direction of the transection and isolate all files necessary
for transection visualization (Figure 11). This step should not be skipped in order to ensure
proper directionality of transection visualization. The two inputs for this model are; the

36

transection’s layer file, and a symbology layer (included in the ETT zipped folder) to correctly
designate the direction of travel along the transection. The outputs generated are a new layer file
of the visualized transection and a new feature class for the transection with additional fields
containing directional information for the creation of the network route in Model 4.

Figure 11. ModelBuilder view of Model 2.

37

3.4.3. FlipTransection 1.2
This model is used to reverse the direction of the transection after it is visualized in step 2 of the
Ethington Transection Toolbox (Figure 12). This step is optional and only required if the user
desires the order of the polygons to go in the opposite direction as those visualized in step 2. This
model may run any multiple of times without complications, however as it only switches the
direction back and forth, this should not be necessary. If the direction of the line is already
correct proceed to step 4. The two inputs for this model are; the transection’s layer file, and a
symbology layer (included in the ETT zipped folder) to correctly designate the direction of travel
along the transection. The outputs generated are a revised layer file of the visualized transection.

Figure 12. ModelBuilder view of Model 3

38

3.4.4. GeoprocessingTasks 3.0 Test 6.5
Reported herein are the primary geoprocessing tasks to automate the creation of Ethington
Transections within ArcMap (Figure 13a & b). This model performs a selection of polygons
adjacent to the transection and exports them to a new polygon feature class in the geodatabase.
Then the transection is divided into segments at the junctions with the polygons in order to
determine relative distances traveled along the transection within each polygon. Then center
point along the transection segment within each polygon is determined. Using a network dataset,
a new one-way network route is calculated which numbers the transection segment centroids
according to their order along the transection. Next the new polygon feature class of adjacent
areal units is spatially joined with the ordered transection segments. Lastly, the order of the
polygons is visualized as a layer file. This step requires the use of the ArcGIS Network Analyst
extension version 10.0 or higher, in order to determine the correct polygon sequence. The inputs
for this model are; the transection’s feature class, the polygon feature class, and a symbology
layer (included in the ETT zipped folder) to visualize the order of the polygons. The outputs
generated are a new layer file of the visualized polygon order and additional columns in the
polygon feature class designating their order along the transection and relative distances traveled
within the polygons along the transect.

39

Figure 13a. ModelBuilder view of Model 4.

40

Figure 14b. ModelBuilder view of Model 4.

3.4.5. Organize Polygon Data and Export Table
This model organizes the polygon data generated in “4. Geoprocessing Tasks 3.0 test 6.5” for
export to Excel and graphing in ArcGIS (Figure 15). This allows for quick visualization of the
data in subsequent models. The only input for this model is the joined polygon feature class. The

41

outputs generated are a revised polygon feature class and an Excel table of the joined polygon
data.

Figure 15. ModelBuilder view of Model 5.

3.4.6. Visualize Polygons
Step 6 of the Ethington Transection Toolbox. This tool is intended to show the sequence order of
the polygons to ensure proper placement prior to graphing the data (Figure 16). The two inputs
for this model are; the polygon feature class, and a symbology layer (included in the ETT zipped
folder) to correctly designate the order of the polygons along the transection. The outputs
generated are a new layer file visualizing the polygon sequence via labeling using the “Order”
variable.

42

Figure 16. ModelBuilder view of Model 6.

3.4.7. Transection Graph_Horizontal (or _Vertical)
This model is to create a line graph of the data from the execution of previous steps in the
Ethington Transection Toolbox and exports that graph in a designated format. There are two
versions of this tool. The first one shown in Figure 17 will display a graph oriented in the
horizontal direction. However if the orientation of the chosen geographic data is more vertical,
the alternate version “7. TransectionGraph_Vertical” should be utilized. The orientation of the
graph directs the choice of plot type chosen to create the template for the model, therefore a
separate template was needed for each graph orientation. Only one template can be specified at a
time in a model, therefore two separate models were needed to allow for both directional
eventualities. The two inputs for this model are; polygon feature class, and a graph template
(included in the ETT zipped folder) to visualize the data plot in a similar style to that of the
manually created Ethington Transections. The output generated is a new graph file.

43

Figure 17. ModelBuilder view of 7. TransectionGraph_Horizontal

In summary, the final version of the ETT only requires two novel data inputs from the
user. All other parameters set by the user while running the ETT are required decisions being
made by the user in order to customize the end result. The primary output as the end result is a
layout consisting of a graphically stylized map and a line graph.
3.5 Documentation
The documentation process occurred throughout application development in order to maintain an
organized experience for both the developer’s programming and the end user’s visualization. The
general methods of documentation are listed in Figure 9 but are discussed in greater detail here.
The first means of documentation was within each model diagram to label major steps and to
identify the specifics of each parameter inside ModelBuilder’s edit window (Figure 10 through
Figure 17). This aided in application development as modules were moved and edited in order to
fine tune the results of the model. The second means of documentation was to construct context
sensitive help that appears during the running of the model if the help window is open. A

44

summary entry was written for each model and then for each parameter from ArcCatalog’s
information edit dialog. Each of these documents have been saved as PDFs and can be reviewed
in Appendix A. Next, a User’s Manual was created in Microsoft Word and exported as a PDF for
cross-platform capability (Appendix B). This document contains step-by-step instructions for
the entire process, especially emphasizing steps needed in between the running of the sequential
models. In order to supplement the User’s Manual and create the most straight forward user’s
experience, short instructional “how to” videos corresponding to each model were produced.
These are video screen captures of the models being run along with narration describing the steps
and parameter settings to use the tool effectively. The videos were generated with TechSmith
Jing and their hyperlinks were integrated into the User’s Manual (Appendix B).
3.6 ETT Evaluation
During the course of the coding process, all models were periodically evaluated for individual
effectiveness by running the ETT multiple times using different transections whose intermediate
data outputs had been previously determined manually. The tests were all run on the same data
that rendered the initial Ethington Transections compiled manually. This data was stored in a
file geodatabase that was amended during the testing phase to include the sample transections
and intermediate outputs. There were three different transections (Error! Reference source not
ound.) that were created as tests and all were line features in one feature class called “Sample
Transections” that were selected by various methods (select by attribute, select by geometry and
using the ArcMap interactive selection tool) between Model 1 and Model 2. The three test
transections used to evaluate tool performance during ETT application development
(SimpleLine, Spiral Track and Tracks) can be seen in Figure 18 along with the feature classes
visible in the catalog window on the right side of the figure within the labeled dataset. The ETT

45

generated outputs were then compared to the expected outcomes. When discrepancies occurred
the programming was altered until an adequate solution was achieved (“debugged”).

Figure 18. The three test transections used to evaluate model performance.

Three times during the programming process, two members of the thesis committee also
evaluated the ETT for ease of use and generation of understandable outputs. One committee
member would be considered an expert/experienced Esri ArcGIS user, who teaches graduate
classes in the spatial sciences that often utilize ArcGIS. The other member is a professor outside
of spatial sciences, but who utilizes ArcGIS in his work and would be considered an intermediate
level user. These two differing levels of experience helped to evaluate the functionality and ease
of use of the ETT. The models were distributed to the committee in compressed folders via a
cloud-based based file sharing and storage platform that was then downloaded for testing onto
their ArcGIS system. In general, to successfully guide the committee members in running the
most current versions of the models, the developer relied on creating context sensitive help (that
were also output as standalone pdf documents; Appendix A), detailed user’s documents
(Appendix B) at each step of testing, along with run-through videos detailing each step of the

46

process to be tested. Occasionally the developer met with the committee members and/or Esri
tech support in a virtual web conference room to go through and discuss pending concerns. All
of this valuable feedback was then used to further refine the ETT. However, the committee
members did not compare these outputs to expected results.
3.7 ETT Efficacy Test
In order to evaluate the efficacy of the ETT three different tests were performed. The first was to
compare the outputs of the ETT to the same visualizations Ethington (Forthcoming) created of
streets through Los Angeles County using the historical manual process. The next test was also
performed in urban environment on a different means of transportation—Washington DC’s
Metro. And the final test endeavored to broaden the scope of the ETT beyond urban history by
evaluating shifts in the Environmental Sensitivity Index (ESI) as one travel’s through the Florida
Intercoastal Waterway (ICW).
3.7.1 Old vs. New: Comparison of Pico-Whittier Transections
The ETT was run using the same original geodatabase Ethington that generated Figure 3. This
particular transection was chosen because it was deemed as the most appropriate prototype for
the first iteration of the ETT and thus was deemed the most applicable first test. One input was
the polygon feature class that contained the percent of whites per census tract in Los Angeles
County for years 1940, 1960, 1980 and 2000, while the other was the street line feature class.
The geographic data is oriented in the horizontal, therefore it was chosen direction for the final
model. The tool was run once through for models one to six, however model seven was run four
different times to generate the graphs based on four distinct time periods. All graphs were saved
and then added to the layout from the graph manager.

47

3.7.2 Washington DC Orange Metrorail
The next transection created was of the Washington DC’s Orange Metrorail as it traverses the
census tracts of the city. The ETT was run using the data acquired from the District of Columbia
Open Data Catalog (DCGISopendata 2015a, DCGISopendata 2015b). This particular transection
was chosen because it was similar in nature to the original Ethington transections by being both a
well-defined urban environment and of a means of transportation, however it was a slightly
different take on both of these components. The public availability of the data was also a benefit
for the project to allow for easy acquisition at no cost. The Orange line was chosen because of
its primarily horizontal orientation and the percentage of the line that occurs inside the borders of
Washington DC as opposed to neighboring VA or MD which is beyond the bounds of the
visualization. One input was the polygon shapefile that contained the demographics of race and
median household income per census tract in Washington DC for 2000 and change since 1990,
while the other was the Metrorail line shapefile. The horizontal graph was chosen for the final
model. The ETT was run once through for models one to six, however model seven was run
three different times to generate the graphs based on three different variables. All graphs were
saved and then added to the layout from the graph manager.
3.7.3 Environmental Sensitivity Index along the Intercoastal Waterway
The last transection created was by analyzing how the ESI changes along the course of the
eastern Florida ICW. The ETT was run using the ICW data acquired from the Florida Fish and
Wildlife Conservation Commission (FWC-FWRI 2001) and ESI data from the National Oceanic
and Atmospheric Administration (NOAA 1996). In particular the polygon feature class
representing the number of monitored rare species of reptiles was chosen due to the nature of the
polygons in additional categories that will be discussed further in Chapters 4 and 5. This

48

particular transection was chosen because it was so different from the first two, and to showcase
the broad application of this technique beyond that of the urban historian. The public availability
of the data was also a benefit for the project to allow for easy acquisition at no cost. The vertical
orientation of ICW along Florida’s eastern coast is also an opportunity to display a contrasting
vertical transection. One input was the polygon geodatabase (GDB) that contained the ESI values
for all of peninsular Florida in 1996, while the other was the Eastern Florida ICW line shapefile
(FWC-FWRI 2001). The vertical graph was chosen for the final model. The ETT was run once
through for all models. The graph was saved and then added to the layout from the graph
manager.
3.8 Application Development Issues Encountered
Several issues were encountered during development of the Ethington Transection Toolbox.
Early in the process of creating sample transections, it was discovered that paths that “turned
back on” themselves could not render geographies that line up visually with the graphed data.
Although this is a logical conclusion in retrospect, it made for some interesting investigations
into how to handle this particular situation. In the end, the decision was to accept this limitation
and document it in the usage of the ETT.
In the first model it was discovered that even though the coordinate system was defined
for the feature classes copied into the newly created feature dataset, the feature data set’s
coordinate system was not defined. This is contrary to Esri documentation. An Esri support
personnel was questioned and unaware of the discrepancy. Esri made note of it for future
exploration. Consequently a “Define Projection” module was added and exposed as a parameter
to ensure that a proper coordinate system and projection is defined from the outset for the feature
dataset and its future feature classes.

49

It became apparent that the direction of the transection could not always be predicted
from an inspection of the data and that the user needed to be aware of and have control over the
direction of data display and polygon order. Most graphs are either interpreted from left to right
or top to bottom and if the polygon order was in fact opposite of this convention the data would
not align with the geography. Therefore; a way to visually inspect the direction of travel along
the transection was needed. Model 2was created for this purpose. All line features in ArcMap
contain directional information that is utilized in network analysis, but it is not a data attribute
that is accessible under normal circumstances. A symbology layer with arrows denoting the
directionality of the line was created and applied to a feature layer based on the transection
feature class, and a field is added to the attribute table indicating the direction of travel. This
attribute can then be inspected and altered later if needed. The user has the option of accepting
the current direction of transection or to run Model 3 to reverse the transection’s direction. The
modules would not visualize properly when combined with Model 1 and thus remained in a
standalone model. This addresses the most pervasive issue of ETT’s application development.
The most challenging issue arose due to the size and complexity of the primary model
(mostly encompassed in Model 4). As additional functionality was added to the model and the
amount of modules increased, certain necessary intermediate outputs were not generated, nor
added to the map despite their designations as parameters and the selection of “add to display”
options. This turned into a recurring phenomenon throughout the programming process. For
instance once the buffer process was added to the model the polygons would no longer appear on
the map and the labels for the order were not being applied from the symbology layer.
Additionally when the visualization modules were run concurrently with the export of the table
to Excel, neither worked successfully. Both of these steps became separate models (5 and 6).

50

When these geoprocessing tasks were run outside ModelBuilder the visualizations were as
expected. However copying the geoprocessing tasks from the results window and creating a new
model resulted in lack of visualization again. Breaking the model apart and running the pieces
separately worked. Combining them back together did not. Esri support was contacted and they
worked to combine the models and to generate the intermediate outputs, but were also
unsuccessful. The recurring solution was to break the model apart into more models.
Consequently the tool was divided into seven models, rather than only three models
Another issue was discovered when the selection of the polygons along the transection
first occurred as many adjacent polygons were not included in the selection. The precision of the
selection tool and that of the copy was not as fine as expected. Therefore a buffer was used to
ensure that all desired polygons were included in the selections. The size of the buffer is a
parameter in the model and can be user adjusted to an appropriate size for the scale of the
visualization. The same buffer was also necessary to select all the transection lines divided at the
polygon borders. It seems that once the lines were divided they did not always fall directly on
the original transection. However, at the scale of Los Angeles County this buffer was only 10 m
wide and therefore sufficient to include the appropriate line pieces without selecting additional
pieces.
There was no way to automate the creation of a new network dataset based on the
transection. This particular issue was well documented by Esri. The only concession to this
particular issue was to ensure that the steps necessary to create the network data set were well
documented in the User’s Manual and the corresponding videos. Additionally due to the nature
of network datasets and the creation of routes, it was discovered that the line features utilized for
the transection have to be contiguous (no gaps) for a successful run.

51

The last model of the ETT was not without its challenges either. It turns out that although
the creation of graphs can be automated in ModelBuilder (or Python) the placing of graphs onto
the layout cannot. Therefore the last steps of the process are well documented in the User’s
Manual and videos.
In summary, although there were several issues encountered during the application
development process, solutions were implemented as best as possible. The primary responses
turned out to be splitting the models into smaller components and increasing the level of
documentation associated with that process.

52

CHAPTER 4: RESULTS
The primary results of this project are the ETT itself and the outputs generated from running the
models. The functionality of the ETT was previously discussed in “Chapter 3. Methodology,”
thus this chapter will be devoted to discussing the different outputs produced throughout the
process and then the results of the various efficacy tests.
4.1 Running the ETT
The actual step by step process of going through using the ETT is documented in detail in the
context sensitive help (Appendix A) and in the User’s Manual (Appendix B), which is the main
documentation provided for the ETT, developed as part of this thesis work. A summary of the
procedures for using the ETT is provided in this chapter. The ETT comes packaged in zipped
folder along with supporting files and sample data (Figure 19). Once these files are copied into
the desired working folder the ETT can be opened up in the Catalog window of ArcMap (Figure
4).

Figure 19. Contents of Zipped folder for delivery of the ETT.

In order to start a visualization, ArcMap is opened to a blank map with the Catalog
window open to the ETT. All of the ETT models, except for “6. Visualize Transection,” are run

53

by double clicking on the model’s icon in the Catalog window. After running the first model the
ArcMap windows will look similar to Figure 20, where the polygon and line data is added to the
map after being copied into a new feature dataset inside a new file geodatabase inside a new
folder (indicated in the red box in the catalog window on the right).

Figure 20. ArcMap after model “1. Transection File Set Up” runs.

The next model isolates the contiguous line features for the transection and then
visualizes the direction of travel along the line (Figure 21). For example, a new feature class
entitled “Transection_Dissolve” is created and stored in the GDB.

54

Figure 21. ArcMap after Model 2 runs.

The only visible change after running Model 3 is that the direction of the arrows has
flipped (Figure 22). However since this is not the desired direction for graphing, the tool is run
again to flip the direction back again to a West to East orientation before proceeding in the
sequence. In general most graphs are produced and read with the starting point on the left-hand
side. This recognition pattern correlates with a path that travels from the left to the right instead
of the right to left orientation of Figure 21. After the network dataset was created, prior to
running model 4, and can be seen in Figure 23.

55

Figure 22. ArcMap after Model 3 runs.

Figure 23. Example of how the ArcMap windows prior to running Model 4.

56

Figure 24. An example of ArcMap after Model 4 runs.

As the sequence continues the polygons are selected and saved as a new feature class
(Figure 24). These polygons have the information for the variables of interest as well as the
sequential designations along the transection. These new files highlighted in the catalog window
by the red box. And running model 5 generates an Excel table from the joined polygon attribute
table as seen indicated by the red box in the catalog window in Figure 25.

Figure 25. An example of ArcMap after Model 5 runs.

57

Figure 26. An example of ArcMap after Model 6 runs.

In the last few steps the final the transection visualization is accomplished. In model 6 the
geography is labeled and situated with respect to the transection (Figure 26). And finally the
graph is produced and placed within the layout, and sized to match the length of the geometry
(Figure 27).

Figure 27. An example of a final layout view for a sample ETT run.

58

The total time completely run through the ETT is between 30 and 45 minutes depending
on how much time is spent in between running the models making adjustments or saving
intermediate data.
4.2 ETT Efficacy Test Results
The Efficacy tests on the ETT allowed for comparisons with known “Ethington Transections” as
well as a chance to showcase new possible applications. While these goals were accomplished,
each new test also managed to illustrate some of the limitations of the current version. These
limitations are detailed below with coordinating figures to best explain the necessary
assumptions of the ETT’s use and how the user can navigate these particular speed bumps while
running it.
4.2.1 Old vs. New: Comparison of Pico-Whittier Transections
Two things became readily apparent when attempting to run a replica of the Pico-Whittier
Transection in Figure 3. The first was that there was a large break between the two streets as
they did not meet exactly in the middle. Since a primary requirement of using the ETT is that the
transection be contiguous, the two sides could not be run simultaneously. Pico Blvd. was chosen
as the first half of the graph. Once Pico Blvd. was isolated and the used as the transection for the
first two tools, a second issue presented itself as can be seen in Figure 28—Pico Blvd is actually
shown as going two directions at the same time. This dual directionality would not produce a
valid route in Model 4, and therefore was investigated visually by zooming in on one of the
conflicting areas (Figure 29).

59

Figure 28. Pico Blvd. transection shown going in 2 different directions.
This smaller scale revealed that the line feature class used for Pico Blvd occasionally split
in two, presumably to represent a median or other split roadway situation (Figure 29). Similar
“surprise” situations are likely to occur when using new data sources for the first time with the
ETT. This spilt line was remedied by editing a copy of the line feature class that only contained
Pico Blvd to remove the southern portions of the double lines. The southern portions were
chosen because it was evident that Pico Blvd. ran between adjacent polygons and was coincident
with many of the boundary edges, but in general the line feature class tended run slightly to the
south of the demarcation. It was hoped that maintaining the northern path would allow for more
generous polygon inclusion during the selection process.

Figure 29. Zooming in on Pico Blvd. to reveal split lanes along travel path.

60

Figure 30. Pico Blvd. Transection in only one direction.

After the editing of the Pico Blvd. feature class, the first entire ETT was run through
yielding a single directionality visualized after Model 2 in Figure 30, and the polygon selection
in peach shown in Figure 31. This visualization yielded far too few polygons in comparison to
the original transection. As alluded to previously, the boundary of many of the polygons was
coincident with Pico Blvd as can be seen more clearly the small scale map on the right in Figure
31. The solution to this issue was to enlarge the buffer used in Model 4 from the default 20
meters to 100 meters. This incorporated the additional green polygons in Figure 31’s left pane.

Figure 31. The two alternate polygon selections of Pico Blvd. with a portion in small scale.

After ensuring the proper polygons were incorporated in the final polygon feature class,
the final visualizations were produced in two map documents (Figure C- 1 and Figure C- 2). The
graphs for the percentage of whites in the total population in 1940 and 1960 are on one layout,

61

while the 1980 and 2000 are on another. The graphs that result from using the default actions of
the ETT are stacked and aligned as best as possible according to the directions in the User’s
Manual.
4.2.2 Washington DC Orange Metro Line
Using the ETT to create an Ethington Transection of the Washington DC Orange Metro Line had
some similar issues to the Pico Blvd. test, along with a few novel realizations arising from
utilizing this data set. One similar conundrum was from the phenomenon when there are
coincident polygon boundaries. The selected polygons and paired ordered stops are displayed in
Figure 32, where the transection can be seen traveling on top of adjacent polygon boundaries.
This results in some of the stops along the route being placed in approximately the same plane
(stops 10-13, 18 &19, 29 & 30 and 33 & 34) for the graph. The solved route still distinguishes a
sequential order, however the distance between the grouped values on the graph is small (0 m,
99 m, 5 m, 104 m) relative to the total length of 14,620 m.

Figure 32. The ordered stops displayed along with the selected polygons for the DC test.
An additional evaluation of stops 10 through 13 illustrate what happens when the
transection takes a 90° turn and the logical sequential order no longer is represented adequately

62

by the ETT generated distance on the graph (actually places stop 11 and 12 in the same location
because the distance to the graphing line between them is negative which defaults to zero to
avoid errors). Next a whole can be observed under stop 7, where a polygon completely
surrounded by selected polygons remains unselected since the transection does not course
through it. And further note that several polygons are actually crossed by the transection in
multiple places which may also yield unexpected polygon sequences.
A novel realization occurred upon discovery of the fact that even if both source data sets
were in the same coordinate system and it was specified in the “define projection” parameter the
ETT would yield no results if this was not a projected coordinate system. Both the Metro Line
and 2000 Census Tract feature classes, seen in Figure 33, shared the WGS 1984 geographic
coordinate system. This was discovered when the outputs from Model 4 appeared as new feature
classes within the GDB, but were not seen in the map. Upon “zooming to layer” the Transection
Points were seen in a completely shifted part of the world. The coordinate system was now
designated as “NAD_1983_State_Plane_California_V_FIPS_0405.” This is contrary to Esri’s
documentation that specifically says “when you specify the coordinate system, the network can
be accurately reprojected to match the projection of other map layers that might be displayed in a
common view” (Esri 2015d, “More Information para 1).

63

Figure 33. The DC Metro Line and 2000 Census Tract data.

It is presumed that California State Plane projection was chosen either because it was
listed as the first choice in the favorites list, or because it was the last projection coordinate
system used (with the Los Angeles County data for the Pico test). To alleviate the issue, the two
initial data sources were both projected to WGS_1984_UTM_Zone 18N prior to running the tool
again. No further complications arose due to this oddity during the remainder of the test.
The final layout of this Ethington Transection can be seen in Figure C- 3. Just as with the
DC test Transection, two graphs were stacked below the map in accordance with the User’s
Manual instructions. The two variables that were graphed, Percent of Total Population Under 17
and Percent of Owner Occupied Homes, were chosen because the normalized values had been
already calculated in the source data and so were the most efficient choices for the visualization.
4.2.3 Environmental Sensitivity Index along the Intercoastal Waterway
Executing the ETT on the ICW and ESI proved again to be a circumstance where the transection
data source had to be altered prior to a successful run. Although the ICW feature class was a line

64

type as required, the visualization in Model 2 revealed a discontinuous multi-directional split
path. When inspected in a smaller scale, as in Figure 34, it could be determined that the line
features were shaped as the edges of a long thin polygon defining the two edges of the navigable
channel as the ICW snakes its way along the coast. Furthermore there were two instances where
the polyline feature endpoints were not adjacent, resulting in a broken topology. Editing a copy
of the source file remedied the situation.

Figure 34. The initial line feature class of the ICW.

The ESI GDB contains several possible categorical feature classes giving counts the
numbers of rare species found in certain polygonal areas that were candidates for using with the
ETT. However before using the ETT, these feature classes were visually inspected with regards
to their spatial and attribute qualities. Several different polygon arrangements were discovered
that required some consideration; overlapping features, multi-part features, a discontinuous
polygonal fabric, polygons crossed by the transection multiple times, and corridors narrowly
sheathing the transection or inversely creating an exclusionary buffer around the transection (all
of which are represented with further explanation in Table 3). The reptile feature class proved to

65

be the most readily appropriate for running the ETT and it can be seen mapped along with the
southward flowing ICW in Figure 35.
The final Transection can be seen in Figure C- 4 where the north-to-south orientation of
the layout necessitated the vertical graphing template. Both orientations of graphing template
were created with the expectation of normalized percent data and so the scale was set to range
from 0 to 1 with a grid line indicating the 0.5 mid-point. Since these data were raw species
counts the initial graph rendered a seemingly bizarre set of lines that went to zero and off the
edge of the graph again. Adjustments to the scale were then made to display the full range of
values by accessing the graph’s properties while it was open in the graph window prior to adding
it to the map layout. Even with scalar modifications, the large shifts in species counts between
polygons (from 0 to 285 in some cases) and a preponderance of zero values still gives a
stochastic, rather than a fluid feel to the Transection.

Figure 35. The reptile feature class of the ESI with the Floridian ICW.

66

CHAPTER 5: CONCLUSIONS AND DISCUSSION
The goal of this project was to create a more efficient and user-friendly way to conduct an
Ethington Transection visualization. By comparing the time required to historically conduct the
visualization (8-10 hours) to that of using the ETT (30-45 minutes) it can be concluded that the
ETT is more efficient. However there are several considerations and limitations in the use of the
ETT, and they are detailed in this chapter. And although some limitations may eventually prove
to be permanent constraints, it is supposed that in the future, some solutions to these limitations
may be found and incorporated into the ETT. Additionally, there is still a great deal of room for
improvement to make the ETT more user-friendly and visually appealing.
5.1 Considerations for Using the ETT
There are some considerations for employing the current version of the ETT stemming primarily
from three things; the nature of an Ethington Transection, the limits and inconsistencies of
ArcGIS, and the constraints on the developer’s time and abilities as undertaken for this thesis
project. Some are discussed here to assist future users in their decision making efforts before
and during the use of the ETT.
One of the fist items a user must consider is the nature of the transection they are
planning to use. Certain line feature class qualities make the use of the ETT problematic or
impossible; such as when a transection is discontinuous, bi-directional, or coincident with
polygon borders. A further elucidation occurs in Table 2 of figures elsewhere in the document
that illustrate this concept, the situations when these might occur, what current work-arounds
enable their use as Ethington Transections and whether they can still be used with the ETT as a
transection (column ETT OK?).

67

Table 2. Qualities of a line feature Class that warrant consideration when using the ETT.
Quality Situation it occurs Current work around ETT OK?
Discontinuous
network
(topology error)
Figure 34
When the line feature class
is supposed to be
continuous but drawing
errors produced
unintentional gaps
Repair the topology of the
network by editing out the
gaps in the source data
Yes, if fixed
Discontinuous
network
(by design)
Figure 3
Where two intentionally
separate lines are desired
to be combined in the
visualization to represent
one transection, such as in
Ethington’s original Pico-
Whittier Transection
Visualize the different
sections separately and
then unite in final layout or
combine the exported tables
to create a combined graph
with the data and create
one map for faster version
than if done by hand
Not really,
but ETT can
still speed
time to
visualization
by
automating
many of the
key steps
Bi-directional
Figure 28, Figure 29 &
Figure 34
When two parallel lines are
used to represent some
quality in the transection,
such as medians in the Pico
Blvd. or the channel edges
of the ICW
Edit the source data such
that only one continuous
line remains, the tool will
resolve any directional
inconsistencies
Yes, if fixed
Coincident with
Polygon Boundaries
Figure 31
When the boundaries of the
polygons used correlates
with the transection in some
way and results in both
following the same path
Ensure buffer is sufficient to
include both “sides” of the
transection and may need
to make decisions about
data placement along the
graph line
Yes
Contains Angles
approaching 90° or
less
Figure 33
When the line bends to
such an extent that the
angle approaches 90° it can
causes multiple stops on
originating in different
polygons along the
transection to pile up on
area of the graph line; acute
angles can cause the
sequence of the stops to
actually visualize in a
reverse order
For angles approaching 90°
sometimes stop points can
be manually moved to
provide a measure of visible
distance between graphed
values, but this becomes
less available at 90° and
less, the only available
strategy would be to
average the values of all
“stacked” points and display
on one point along the
graph line
Minimally for
angles
greater than
90°, but less
so without
manual
recalculations
for acute
angles

After the transection, it is wise to consider the nature of the polygons. Table 3 organizes
the primary polygon qualities of concern along with figurative examples and brief descriptions of

68

the matching considerations. For instance, sometimes census data uses your “transection” (i.e.
street, river, or railroad) to demarcate the boundary for the tract polygons. This can make for
challenges in determining the proper way to visualize polygon data that co-occur on either side
of the transection. Are the two values averaged and displayed as one point on the graph? Or
possibly is one of the points (stops) assigned to the polygons manually moved “up or down”
along the transection so they are visualized as distinct values? (Consequently these same possible
strategies can be employed when transection angles are 90 degrees or less, or overlapping
polygons occur.)

Table 3. Qualities of polygons that warrant consideration when using the ETT.
Figurative Example Quality Necessary Considerations

Transection passes through boundaries
multiple times
ETT will designate more than one
location for this polygon which must be
deleted prior to graphing once the user
determines which location is the most
appropriate for plotting the data

Overlapping polygons

May make interpreting transection difficult
as full extent of all polygons will not be
visible in the map portion and so relating
data to location becomes problematic
Discontinuous polygonal fabric Any data gaps are not indicated in the
graph as all data points are connected via
a continuous plot line; the separated
polygons in the graph will appear as the
data seamlessly transitions between
them although actual values are unknown

Polygons have coincident boundaries
with Transection
User will need to decide to what extent to
incorporate “both sides” of the transection
and if these values will be plotted as an
average, two slightly separated values
with an artificial distance, or some other
mechanism when the center locations
along the transection are coincident

69

Figurative Example Quality Necessary Considerations

Multipart polygon features ETT will plot same value in multiple
locations along the transection; these
values may truly represent identical
situations in alternate locations or the
value may be an average of all parts; a
user may decide to leave these results as
the default, plotted at only one location or
to separate them based on their
knowledge of the data’s distribution.

Polygon creates ensheathing corridor
around Transection
ETT handles this expected though it may
appear as if there is a gap in the
polygons at the large scale of the map
graphic; only caution is to not interpret
data values beyond the corridor, which is
the case for all Ethington Transections

An polygonal exclusionary buffer exists
as a corridor around Transection
No polygons would be selected and no
data plotted unless the buffer was made
large enough to account for this
phenomenon; care must be made when
inferring that the conditions of the
buffered polygons accurately represents
the reality inside the corridor

There is no right or wrong choice of strategy in any given situation, but the user must
confront the challenge and make knowledgeable decisions about how to proceed. There indeed
may not be a one-size fits all solution to these challenges, or even a “correct” consistent strategy
employed to address the same quality challenge in subsequent visualizations. It is simply that
these things must be carefully considered and actively addressed during the Ethington
Transection creation process and in particular if employing the ETT.
The next consideration applies to the buffer employed in Model 4 to effect the number of
polygons included in the visualization. The buffering step was originally added to the model to
account for inconsistencies in ArcMap’s drawing of the individual line pieces of transection
inside each polygon. The center point of each of these lines is used to locate each stop in the
sequence and thus determine the order and relative distance on the graph. However once testing

70

of the ETT began, an insidious use was discovered. The raising of the value beyond the level
needed to account for redraw inconsistencies included not merely adjacent polygons, but ones
that were only nearby and within the buffer’s range.
These polygons would not automatically be included in the final visualization if the
transection does not pass through them and the User’s Manual instructions were followed
exactly, however by manually entering an appropriate sequence number for these additional
polygons, and editing the original polygonal attributes in kind, the tool could be wrought to
maintain them in the final visualization. This workaround is surprisingly simple to include
additional polygons. The removal of any unwanted polygons from the increased buffer selection
is even easier. Just delete these features from the resultant attribute table at the same point in the
ETT, between Models 4 and 5. The reverse cannot be said. Manually adding any unselected
polygons at this point involves hunting down the entire records in the original attribute table and
copying and pasting feature by feature. This procedure is cumbersome, time-consuming and
fiddly. Therefore it would seem wise to err on the side of polygon over-inclusion in the first
iterations of any visualizations with the ETT. Consequently manipulation of the buffer can be
used to address the MAUP of Ethington Transections.
5.1.1 How the MAUP applies to Ethington Transections
The Modifiable Areal Unit Problem (MAUP) presents itself in Ethington Transections and when
using the ETT. MAUP arises because the data is displayed at an aggregation level greater than
that at which the data was collected and often “arbitrary with respect to phenomenon under
investigation” (O’Sullivan and Unwin 2010, 37). The whole point generating an Ethington
Transection is to change the spatial reference frame, which in so doing O’Sullivan and Unwin
(2010, 38) imply becomes “itself a significant determinant of the …patterns we observe.” The

71

primary consideration here is the duality whereby the data plotted from the transection line may
not accurately represent the qualities of the whole polygon area shown in the map (a MAUP
typical of most polygon data distributions), or conversely the aggregated data of the whole
polygon may not represent the realities experienced while traveling along the narrowed corridor
of the transection in a “street side view.” It follows then that the MAUP is inherent in any
Ethington Transection, so no matter good the automation of future ETT versions gets, the MAUP
will persist. Users should utilize their knowledge of the subject to evaluate how the phenomenon
they are trying to visualize may be influenced by any aggregation that occurred compiling the
candidate polygons and how increasing or decreasing the number of polygons included in the
process may skew representations of the data. The quantity of included polygons can be most
effectively manipulated via the buffer designation in Model 4 of the ETT. The goal of any user
should be to minimize the extent of the MAUP as much as possible when creating Ethington
Transections and acknowledge it during any discussion during their presentation.
5.2 Limitations of Current ETT
One of the first encountered limitations was that the path or line used for a transection cannot
deviate more than 90 degrees along its course, or to turn back in on itself, in order for the
geography to line up with the data in the graph. The ETT can still be used in this circumstance
but will create a graph where the data will be spaced along the axis such that it represents the
distance traveled along the transection. This was deemed an appropriate compromise and is
likely to remain a constraint in the future.
As briefly discussed previously, another limitation is that the line feature used as the
transection must be contiguous. This is due to the fact that the polygon sequential order is
determined by creating a route using the Esri Network Analyst extension. And although this is a

72

quick and accessible solution, it is speculated that an alternative may present itself in the future
to allow for gaps in the transection and the corresponding graph in future versions of Network
Analyst. Since this was already handled elegantly in Ethington’s visualization of Pico and
Whittier (Figure 3), it is surmised that in future this limitation could be removed from the ETT.
A further limitation imposed by utilizing the network analyst extension is the necessity to
create the network dataset outside the ETT in between models. Whereas this was addressed by
providing detailed documentation for creating the particular network dataset utilized by model 4,
it would be far more elegant a solution if this was also automated. If an alternative method of
sequencing the polygons could be determined, it is likely that these three limitations imposed by
its use could be removed.
The complicated graphing capabilities of ArcMap also presented a challenge to the
development of a user-friendly ETT. There is a great deal of power and flexibility in the
graphing modules of ArcGIS, but they are exposed in such a limited way during the automation
process. All orientation and design decisions (including scale) must be built into the graph
templates and creating a plethora of templates to anticipate all transection graphing eventualities
proved difficult and cumbersome. Ultimately only one horizontal and one vertical template were
provided. The flexibility in the data used for creation of the graph is delivered by exposing the
graphing options as a parameter. However, this parameter appears as one large scrollable table
that can be described as dense at best and opaque to many (Figure 36). The ETT would benefit
from a more simplified parameter, or even a stepped response approach (such as employed in the
graphing “wizard.”
Once the graph is created, there was still no ability to “add to layout” and thus further
steps outside of the ETT were documented to ensure an even remotely comparable transection

73

visualization could be achieved. The graphs can be altered after creation, but this is tedious and
be best handled by allowing graphing options to be available to future programming efforts. It is
desirous in future that more of the graphing functionality could be accessed via programming not
just the creation of multiple templates and a one –size fits all input series.

Figure 36. The Input Series for creating a graph via an exposed parameter in a model 7.

Although the ETT does generate a matched graph and geographic image, it is far from the
elegant and refined transections originally produced “manually” by Dr. Ethington. It would
require a great deal of time to fine-tune the final layouts. As the graphing and feature
manipulation in ArcGIS continues to improve, hopefully the quality of transections produced by
future versions of the ETT will improve as well.

74

The limitations above speak to specific steps in the ETT, however one over-arching goal
of the project was not met and remains a constraint on possible users—that was to make the ETT
truly easy to use and available to those with only a minimal GIS skill level. The multitude of
models and bits and pieces that still necessitate manual completion by the user clearly put the
ETT out of the reach of the novice. It has been determined that an appropriate user would be
considered of an “Intermediate” level. Minimally a user would require knowledge of: using a
ModelBuilder toolbox from both the Catalog and Model Editing window; managing
geodatabases; employing extensions; and utilizing the Catalog window heavily for managing
folders, copying files, interpreting GDB contents, and placing items in the map window. Lastly
users must be able to review all the required parameter inputs and understand the file types and
definitions and how they are applied. The ability to edit input data sources when potential issues
are discovered when using the ETT is also beneficial. It would certainly be desirous to lower the
current skill level needed in the future.
5.2 Addressing Current Issues in Future Work
Part of the reason there was such a rigorous approach to documentation was to counterbalance
the complexity and number of steps necessitated in the final version of ETT. The goal was to
create an easy to use and efficient ETT, but this proved problematic with current programming
constraints. Consequently increased effort went into documenting how a user walks through the
use of the tool in the most straight-forward manner. The addition of the Jing videos integrated
into the User’s Manual proved to be an effective means of achieving this goal. However in
future the ETT would benefit from refinement and a reduction in the total number of models
needed to accomplish the visualization.

75

In the future this issue should be solved so that the number of steps or models required to
run the tool can be reduced. This could be accomplished simply by combining several of the
models. For example models one and two would combine yielding the new GDB and the
proposed transection and direction visualized. Model three should be left separate as an option
when the direction of the transection should be flipped, but models four through six should be
able to be combined in future work with the addition of model seven as well if future Arc GIS
ModelBuilder parameters allow alternate choices of graph templates. This would render only
two or three final models and the optional model three “Flip Transection.”
An additional annoying consequence of breaking the ETT into so many pieces is having
to designate the newly created GDB again in every model, even though it is the same each time.
Additional programming efforts and strategies (such as converting the models to Python) could
fix this issue in future work.
Another future work prospect would be to build-in an ability of the toolbox to batch
process and automatically combine transections based discontinuous networks by design.
Automating this potential situation would increase the potential scope of possible transections
available to the ETT, and further the goal of the developer to be able to generate Ethington
Transections comparable to those of the model “Pico-Whittier Transection” seen in Figure 3.
In order to facilitate future development and use of the ETT, an entire geoprocessing
package is being made available via a public GitHub repository labeled identical to the title of
this thesis. The custom toolbox with all the models, a sample data set, the User’s Manual and
tutorial video files are all offered free to all GitHub users. It is hoped that future versions of the
ETT will benefit from its availability on the web for in the spirit of “open source is the idea that

76

by sharing code, we can make better, more reliable software” (GitHub 2015a, “Use someone
else's project” para 1).
5.3 Additional Future Work
The original vision of the developer was curtailed by the issues and time constraints of this
project. The future iterations of the ETT would benefit from implementing some of the
strategies originally envisioned. The current models could be converted to Python and additional
streamlining by the use of variables could reduce the number of parameter entries necessary.
However this alone might actually raise the difficulty level of running the ETT, depending
entirely on the user’s familiarity with using Python code.

Figure 37. Mock-up of possible future ETT distributed as a Python add-in toolbar.

But if the Python code was then packaged as an ArcGIS add-in, the ETT could begin to
become the interactive easy to use visualization desired. A Python add-in is a customization to
an ArcGIS for Desktop application (that is, ArcMap, ArcCatalog, ArcGlobe, and ArcScene), for
instance a collection of custom tools on a toolbar offering additional functionality (Esri 2015c).
A Python add-in is a single compressed file with an .esriaddin extension delivered with all the
required files necessary for it to work. Add-ins are easy to share between users and plugged into
a desktop application by copying the file to a well-known folder, or simply removed by deleting

77

it from this same folder. The ETT could then be distributed as a toolbar with buttons, menus and
interactive tools that guide the user through the visualization like what is seen in Figure 37.
Because add-ins are also able to make a customization that performs an action in response to an
event, or requires the use of the mouse to interact with the display, additional functionality and
user customization could be incorporated into the ETT. An example might be a tool that allows
the user to interactively click and drag a custom transection path over a map to define an area of
interest, or to limit the visualization area with a drawing tool. Another example might be the
ability to interactively add and subtract polygons from the selection, or alter polygonal sequence
order. A future iteration of the ETT could function as illustrated in Figure 38. All of these
features would greatly enhance the ETT and assist in lowering the skill level requisite for its use.

Figure 38. Mock-up of screen shots while using future versions of the ETT.

78

5.4 Conclusion
This thesis project devised a new method to automate, increase the efficiency, and improve the
efficacy of creating Ethington Transections (Forthcoming), thus hastening time to data
visualization. The final result is custom toolbox called the Ethington Transection Toolbox (ETT).
The previously tedious and time intensive task of manually creating transections was automated
using Esri’s ArcMap, Modelbuilder and Network Analyst, and is now accessible to a wider range
of scholars that will only need a minimal familiarity with ArcGIS to perform this Ethington
Transection visualization. Therefore the ETT should enhance the analytic skills of those with an
interest in urban spatial studies, or other fields, such as environmental studies, that seek to
visualize polygonal data along linear, sequential paths. Automation will enable researchers to
increase the number of transection analyses derived and allow for easier comparisons within and
between places. It is also hoped that other scientific fields will embrace this new visualization
tool for their own purposes and consequently expand the range of ETT users.

79

REFERENCES
Allen, D. W. 2011. Getting to Know ArcGIS ModelBuilder. Esri Press.
Alexander, S., and N. Waters. 2000. The effects of highway transportation corridors on wildlife:
A case study of Banff National Park. Transportation Research Part C: Emerging
Technologies 8 (1–6) (0): 307-20.
Arctur, D., and M. Zeiler. 2010. Designing Geodatabases: Case studies in GIS data modeling.
Second ed. Redlands, CA: ESRI, Inc.
Audisio, C., G. Nigrelli, and G. Lollino. 2009. A GIS tool for historical instability processes data
entry: An approach to hazard management in two Italian alpine river basins. Computers
and Geosciences 35 (8): 1735-47.
Baldwin, S. M. 2014.“Institutional Inscription in the Minorcan Quarter of Saint Augustine,
Florida.” MS thesis, University of Southern California, 2014.
Bell, K. 2004. Automating Work Flows With ModelBuilder at Salt Lake City's Transportation
Division. ArcUser 7(4). Accessed May 7, 2015.
http://www.esri.com/news/arcuser/1104/trafficount.html
Bodenhammer, D. J. 2013. “Beyond GIS: Geospatial Technologies and the Future of History,” In
History and GIS: Epistemologies, considerations and reflections. Eds. von Lünen,
Alexander, and Charles Travis Dordrecht, 1-12; Netherlands: Springer.
Brook, V. 2013. Land of Smoke and Mirrors: A Cultural History of Los Angeles. New
Brunswick, NJ: Rutgers University Press.
Cunfer, G. 2011. The southern Great Plains wind erosion maps of 1936-1937. Agricultural
History 85 (4): 540.

80

DCGISopendata 2015a. Census Tracts by Population – 2000 Last accessed 16 August 2015
http://opendata.dc.gov/datasets/0efdf1f78d4f40e3ba111feadac48553_18
——— 2015b. Metro Lines Last accessed 16 August 2015
http://opendata.dc.gov/datasets/a29b9dbb2f00459db2b0c3c56faca297_106
Dilts, T. E., J. Yang and P. J. Weisberg. 2010. Mapping Riparian Vegetation with Lidar Data:
predicting plat community distribution using height above river and flood height.
ArcUser 14(4): 18-21. Accessed May 7, 2015.
http://www.esri.com/news/arcuser/0110/files/mapping-with-lidar.pdf.
Dunae, P. A., D. J. Lafreniere, J. A. Gilliland, and John S. Lutz. 2013. Dwelling places and
social spaces: Revealing the environments of urban workers in Victoria using historical
GIS. Labour/Le Travail (72): 37.
Esri. 2014. “Esri Insights: Quickly get the facts & demographics for any area.” Esri. Accessed
April 26, 2014.
http://marketplace.arcgis.com/listing.html?id=fef73db0214f4ab9b98b06e645b962f9 and
http://la.arcgis.com/insights/#
———. 2105a. “What is the ArcGIS Network Analyst Extension?” Esri. Accessed August 20,
2015. http://desktop.arcgis.com/en/desktop/latest/guide-books/extensions/network-
analyst/what-is-network-analyst-.htm
———. 2015b. “What is ModelBuilder?” Esri. Accessed August 15, 2015.
http://desktop.arcgis.com/en/desktop/latest/analyze/modelbuilder/what-is-
modelbuilder.htm

81

———. 2015c. “What is a Python add-in?” Esri. Accessed September 5, 2015.
http://desktop.arcgis.com/en/desktop/latest/guide-books/python-addins/what-is-a-python-
add-in.htm
———. 2015d. “24014: The network dataset has an unknown coordinate system.” Esri.
Accessed September 6, 2015. http://desktop.arcgis.com/en/desktop/latest/guide-
books/extensions/network-analyst/24014-unknown-coordinate-system.htm
Ethington, P. J. (Forthcoming). Ghost Metropolis: Los Angeles Since 13,000 BP.
FWC-FWRI (Florida Fish and Wildlife Conservation Commission-Fish and Wildlife Research
Institute) as the data source, and FIND (Florida Inland Navigation District) as the
originator 2001. Intracoastal Waterway East Coast Florida. Last accessed 16 August 2015
at http://geodata.myfwc.com/datasets/facfb0c2903e4c6d89b878b32b3ca610_8
GitHub, Inc. 2015a. “GitHub Help: Fork a Repo” GitHub, Inc. Accessed September 7, 2015.
https://help.github.com/articles/fork-a-repo/
——— 2015b. “Search Results: ArcGIS; Python.” GitHub, Inc. Accessed September 5, 2015.
https://github.com/search?utf8=%E2%9C%93&q=ArcGIS
Gregory, I. and P. S. Ell. 2007. Historical GIS: Technologies, methodologies and scholarship.
Vol. 39. Cambridge, UK: Cambridge University Press.
Lynch, K. 1960. “The image of the environment”. In The image of the city. 11: 1-13. Cambridge,
MA: MIT Press.
Maier, U., M. Flegr, H. Rügner, and P. Grathwohl. 2013. Long-term solute transport and
geochemical equilibria in seepage water and groundwater in a catchment cross section.
Environmental Earth Sciences 69, (2): 429-441.

82

Mishra, R. K., M. Parida, and S. Rangnekar. 2010. Evaluation and analysis of traffic noise along
bus rapid transit system corridor. International Journal of Environmental Science &
Technology 7, (4): 737-750.
Morillas-Torné, M. 2012. Creation of a geo-spatial database to analyse railways in Europe
(1830-2010). A historical GIS approach. Journal of Geographic Information System 4
(2): 176-87.
Monmonier, M. S. 1996. How to lie with maps. Chicago: University of Chicago Press.
Montello, D. R. 2001. “Scale in Geography”. In International Encyclopedia of the Social &
Behavioral Sciences, eds N.J. Smelser and P.B. Baltes, 13501-13504. Pergamon Press.
Murrow, J. L., C. A. Thatcher, F. T. van Manen, and J. D. Clark. 2013. A data-based
conservation planning tool for Florida panthers. Environmental Modeling & Assessment
18 (2): 159-70.
NOAA (National Oceanic and Atmospheric Administration) 1996. Last accessed on 16 August
2015 at http://response.restoration.noaa.gov/maps-and-spatial-data/download-esi-maps-
and-gis-data.html#TOC
Offen, K. 2012. Historical geography II: Digital imaginations. Progress in Human Geography 37
(4): 564-577.
O’Sullivan, D. and D. Unwin. 2010. Geographic Information Analysis. 2
nd
ed. Hoboken, NJ:
John Wiley & Sons.
Pratt, M. 2000. Modeling Fire Hazard. ArcUser 3(3): Special 1-5 Accessed May 7, 2015.
http://www.esri.com/news/arcuser/0700/files/firemodel.pdf.

83

Peterson, G. 2011. Small Multiples: March 30
th
, 2011 blog post. A Cartographer’s Toolkit: map
making tips, tricks and inspiration. Accessed 15 August 2015.
http://www.gretchenpeterson.com/blog/small-multiples/
———. 2012. Cartographer’s Toolkit: Color, typography, patterns. Fort Collins, CO:
PetersonGIS.
Plummer, B. 2014. Data visualization: past vs. present; Blog posted by Ben Plummer October
07, 2014. DataWatch Accessed September 9, 2015. http://www.datawatch.com/data-
visualization-past-vs-present/.
Pulido, L., L. Barraclough and W. Cheng. 2012. A People’s Guide to Los Angeles. Los Angeles:
University of California Press.
Rattan, A., A. Campese and C. Eden. 2012. Modeling Walkability ArcUser 15 (4): 30-33
Accessed May 7, 2015. http://www.esri.com/news/arcuser/0112/files/walkability.pdf.
Raymond, A. 2011. Denny Regrade, 1893-2008: A case study in historical GIS. Social Science
History 35 (4): 571-97.
Reed, S. E., J. L. Boggs, and J. P. Mann. 2012. A GIS tool for modeling anthropogenic noise
propagation in natural ecosystems. Environmental Modelling and Software 37: 1.
Reiser, J. 2014. “Census download tools for ArcGIS”. Accessed April 26, 2014.
https://github.com/RowanGeolab/ArcGISCensusDownload
Roelvink D., A. Reniers, A. van Dongeren, J. van Tiel de Vries, R. McCall, J. Lescinski. 2009.
Modeling storm impacts on beaches, dunes and barrier islands. Coast Engineering
56(11–12):1133–1152

84

Roelvink D., A. Reniers, A. van Dongeren, J. van Tiel de Vries, J. Lescinski, R. McCall. 2010.
XBeach Model description and Manual. UNESCO-IHE Institute for Water Education,
Deltares and Delft University of Tecnhology. Report June, 21 2010 version 6
Silva, A. M. A. N., and R. P. M. Taborda. 2013. Integration of beach hydrodynamic and
morphodynamic modelling in a GIS environment. Journal of Coastal Conservation 17
(2): 201-10.
Slocum, T. A., R. B. McMaster, F. C. Kessler, and H. H. Howard. 2009. Thematic cartography
and geovisualization. Third ed. Upper Saddle River, New Jersey: Pearson-Prentice Hall.
Spalding, V. 2000. Automating Spatial Processes. ArcUser 3(3). Accessed May 7, 2015.
http://www.esri.com/news/arcuser/0700/mbover.html.
Stevens, D., S. Dragicevic, and K. Rothley. 2007. iCity: A GIS–CA modelling tool for urban
planning and decision making. Environmental Modelling and Software 22 (6): 761-73.
SWAN. 2015. Scientific and technical documentation SWAN Cycle III version 41.01A. Delft
University of Technology, Environmental Fluid Mechanics Section. Accessed August 18,
2015 http://swanmodel.sourceforge.net/online_doc/swantech/swantech.html
Tufte, E. R. 1990. Envisioning Information. Cheshire, CT: Graphics Press.
———. 2001. The Visual Display of Quantitative Information. Cheshire, CT: Graphics Press
Tang, Y., and D. W. Wong. 2006. Exploring and visualizing sea ice chart data using java-based
GIS tools. Computers and Geosciences 32 (6): 846-58.
Thomsen, M., T. Wernberg, Peter A. Stæhr, and Morten F. Pedersen. 2006. Spatio-temporal
distribution patterns of the invasive macroalga sargassum muticum within a Danish
sargassum-bed. Helgoland Marine Research 60 (1): 50-8.

85

Vanleeuwe, H. 2010. Predictive mapping of season distributions of large mammals using GIS:
An application to elephants on Mount Kenya. Methods in Ecology and Evolution 1 (2):
212-20.
Watson, R, and D. Pauly. 2014. Coastal catch transects as a tool for studying global fisheries.
Fish and Fisheries 15 (3): 445-55.
Wikimedia Commons 2015. File:Minards chart Napoleons Russian campaign of 1812 made in
1869.jpg. Wikimedia. Accessed September 8, 2015.
https://commons.wikimedia.org/wiki/File:Minards_chart_Napoleons_Russian_campaign
_of_1812_made_in_1869.jpg
Wheaton, Joseph M., Chris Garrard, Kelly Whitehead, and Carol J. Volk. 2012. A simple,
interactive GIS tool for transforming assumed total station surveys to real world
coordinates – the CHaMP transformation tool. Computers and Geosciences 42 : 28-36.
Zandbergen, P. A. 2013. Python Scripting for ArcGIS. Redlands, CA: ESRI, Inc.
Zeiler, M. 1999. Modeling our world: The ESRI guide to geodatabase design. Vol. 40 Redlands,
CA: ESRI, Inc.

86

APPENDIX A: CONTEXT SENSITIVE HELP DOCUMENTS FROM MODELS

87

88

89

90

91

92

93

94

95

96

97

98

99

100

APPENDIX B: USER’S MANUAL FOR THE ETHINGTON TRANSECTION
TOOLBOX (ETT)

101

User Instructions for Ethington Transection Automation Toolbox

1. See “Step1Instructions” video (http://screencast.com/t/pTwBQCJ6kDXy). Copy the zipped folder
“Testing for Ethington Transection Automation Tool”, containing a custom toolbox and all
supporting template layer and graph files, to your computer where you would like to run the
transection automation. (This also includes a copy of the “EthingtonTransection.gdb” for
sample files). You should then unzip the folder where you want to conduct your analyses.

Figure 39. List of all files included in the zipped to run Ethington Transection Automation Tools.
2. Open ArcMap to a Blank Map and choose a default geodatabase. (The choice of default
geodatabase is irrelevant to use of the tool as the first steps will establish a new empty
geodatabase to house all the new files created while running the transection automation.)
3. Open the Catalog window and “pin it” open by pressing the pushpin in the top right corner till it
faces down so the catalog window stays open. Navigate to the folder containing the copied
“Ethington Transection Toolbox” and open it so it reveals Tools 1-7 as below.

Figure 40. Catalog window in ArcMap showing the contents of the toolbox and the supporting files for running the
Ethington Transection Automation.
4. Turn on the Network Analyst Extension by going to Customize/Extensions in ArcMap. Click the
check box for “Network Analyst” (this will be used in Tool 4 and is easier to turn on now rather
than at that time). For the same later use ensure that the Network Analyst toolbar is turned on
by going to Customize/Toolbars/Network Analyst.

102

Figure 41. Extension Manager in ArcMap to Turn on Network Analyst.

5. Double click the tool labeled “1.Transection File SetUp”. When the Tool opens please ensure the
Help is visible by toggling the “Show/Hide Help” button until it is visible to the left of the
parameter input window. Clicking in in each parameter field will change the context sensitive
help messages at the left. By utilizing the information in the help window please make the
appropriate choices for the four parameters and then press OK to run the tool.

Figure 42. The view of Tool 1 prior to running it with parts labeled.
6. Save the Map file to a location of your choice, if you want to come back to this point at a later
date. At this time you should notice that the line and polygon files you selected have been
added to the map and that a new folder, file geodatabase, and dataset with the two files you

103

selected have been added in the catalog window. You will also have to reopen the toolbox at
this time to see the next tool.

Figure 43. ArcMap after Tool 1 runs with the new files indicated in the catalog window.
7. See “Step2Instructions_Alternate” video (http://screencast.com/t/RJDKALGGdUC). Select the
contiguous line features that you want to designate as the transection by using any of ArcMap’s
selection tools (an example selection of four features in one contiguous line selection is shown
in Figure 43). If nothing is selected all features in the line file will be used to create the
transection. If these are not contiguous line features there will be an error in the next step and
when running Tool 4.
8. Run Tool 2 by double-clicking it, reading the help menu items for the tool and each parameter,
filling out the parameter inputs with the appropriate data, and clicking OK.

104

Figure 44. Tool 2 when open showing suggested parameter inputs and help.
9. Save the Map file again and evaluate the direction of the transection indicated in the map
window. The direction of the black arrowheads indicates the direction of travel along the
transection and will indicate the order that the polygon data will be recorded. In general the
direction of travel should be left to right or down to up to coincide with the direction of default
graph reading, but as this can all be customized later it is up to user’s choice. If the current
direction is correct, skip the next step and do not use Tool 3.

Figure 45. ArcMap windows after running Tool 2 with the black arrows indicating the direction of travel along the
transection.

105

10. See “Step3Instructions” video (http://screencast.com/t/t7QqNo8rdaNT).Run Tool 3 if you want
to switch the transection’s direction of travel. Follow the directions in the help and enter the
appropriate parameters.

Figure 46. Dialog of parameter window and help window of Tool 3 with example entries.

Figure 47. ArcMap after running Tool 3 to flip the direction of travel for the transection (notice that the arrows are
pointing the opposite direction than in Figure 45.
11. See “Step4InstructionsBefore” (http://screencast.com/t/xXZ5FrzNOVSe).Create a network route
with the transection. First ensure that the Network Analyst window is open by clicking the

106

button on the Network Analyst toolbar. Move/dock this window where you can access it but it
does not impede the map, catalog or TOC windows.

Figure 48. ArcMap with the Network Analyst window and button indicated.

107

In the catalog window, right-click the dataset and Select a New/Network Dataset as indicated in
Figure 49.

Figure 49. Selecting to create a New Network Dataset by right-clicking on the database in the catalog window.

108

When the Network Dataset wizard begins, enter a name for the dataset (the default is fine), and
press Next. Select only the feature class “Transection_Dissolve” which contains the transection,
and press Next. Answer “No” for “Do you want to model turns in this network?” and press Next.
Leave the Connectivity settings as is and just press Next again. Choose “None” for elevation
models and press Next. The attributes of the dataset should appear as in Figure 50 where you can
optionally set the units of the length attribute (if they do not appear by default) by clicking on
the field at the cross section of “Units and Length”. At the Travel Mode Screen click the + sign
next to the drop down menu as indicated in and Enter any word (I suggest “Going”) and enter.
Then Choose “Not Allowed” from the U-Turns at Junction field. Then press Next. Select “No” to
the question “Do you want to establish driving directions settings for this network dataset?” and
click Next. Leave the “Build Service Area Index” unchecked and click Next. The last screen is the
Summary page and Click “Finish” to create the Network Dataset. Once the wizard runs click
“Yes” when asked if you would like to build the dataset now, but “No” When asked if you want
to add all the elements to the map (lack of spatial references can be ignored). The Network
Dataset files should now be added to the dataset and look similar to Figure 52.

Figure 50. Network Dataset Wizard on attribute screen, with optional setting of units open.

109

Figure 51. Travel Window for creation of new Network Dataset with the two items that must be changed indicated
with red arrows.

Figure 52. Network Dataset Elements identified in the catalog window.
12. See “Step4bInstructions” video (http://screencast.com/t/ch9yrqX1). Run Tool 4 by double-
clicking it and following the help directions for the parameters’ input. This may take a few
minutes to complete.

110

Figure 53. Parameter Entry Window and Help for Tool 4.

13. Add the “Poly_to_Points_SpatialJoin_Sort” feature class from the catalog to the Table of
Contents by dragging the feature class from the catalog window to the TOC window.
14. See “Step5Instructions” video (http://screencast.com/t/uNt4JvZdtl). Designate the Working GDB
in the first parameter input and leave the rest of the parameters as the default (Figure 54).

111

Figure 54. The Parameter Entry Window for Tool 5.
15. See “Step6Instructions” video (http://screencast.com/t/cKTDyoVDT). Run this tool from the
ModelBuilder window by right clicking the tool and choosing “Edit” (Figure 55). The only
parameter input is selected by double-clicking the star at the beginning of the flow diagram
(Figure 56). This should point to the same Working GDB used with all the other tools. Then
validate the tool by clicking the check mark in the upper right hand corner. And run the tool by
clicking the small right facing green triangle in the upper right hand corner. After the tool runs
the labels must be designated. Right click the layer added to the map and select “Properties”
(Figure 57). Select the “Labels” tab and ensure the “Label features in this layer” box is checked,
and that “Sequence” is the designated label field (Figure 58).

112

Figure 55. Selecting the Edit option from the right-click drop-down menu for Tool 6 in order to run the tool in the
ModelBuilder window.

Figure 56. The ModelBuilder window with the steps to run Tool 6 designated in sequential order.

113

Figure 57. Highlighting the Properties option while right-clicking the newly created layer after running Tool 6.

Figure 58. The Labels tab for the Layer's Properties wIndow indicating the correct inputs for properly labeled
transection polygons resulting from running Tool 6.

114

16. See “Step7Instructions” video (http://screencast.com/t/8EFdTML6w). First change to the
“Layout View” (Figure 59). Run Tool 7 in order to generate a scaled graph based on the values of
your transection polygons. First determine whether your map is primarily running in the
horizontal or vertical direction and select the appropriate tool to coordinate. The parameters
necessary for entry are primarily contained in the Input Series table for the graph’s options.
Please use the tool’s context sensitive help to set the desired parameters to generate the graph
(Figure 60).

Figure 59. The two methods to select the "Layout View" prior to running Tool 7.

115

Figure 60. The parameter entry window for tool 7, most of which is taken up by the Input Series for the creation of the
graph.

116

After running Tool 7 you should view your graph by going to the “View menu” indicated in Figure
59 and selecting “Graphs>Load Graph”. Then navigate to the recently saved graph file, select it
and press “Open” (Figure 61). This will open a new window with a graph simply symbolized such
as in Figure 62.

Figure 61. Open menu for loading recently saved graphs to the viewing window.

Figure 62. An example of graph created after running Tool 7, horizontal version.

117

17. The final step is to right-click the graph and choose “Add to Layout” from the menu (Figure 63).
Once the graph appears in the layout view, you must resize the graph bounding box to the same
width as the geography and this will align the results of the graph to the center points of the
polygons to which they correspond. At this point you may generate additional graphs, export
the graphic to a graphics editor for further refinement, or generate your final output from the
layout view (Figure 64).

Figure 63. When right-clicking the graph window a context sensitive menu appears. The "Add to Layout" option is
indicated by the red arrow.

118

Figure 64. An example of the end result of following these instructions for using the Ethington Transection Automation
Toolbox.

119

APPENDIX C: MAP DOCUMENTS PRODUCED DURING EFFICACY TESTS
Figure C- 1. Pico Blvd. Transections for Percent White in 1940 and 1960

120

Figure C- 2. Pico Blvd. Transections for Percent White in 1980 and 2000

121

Figure C- 3. DC Orange Line Metro Transection for Percent <17 and Home Ownership.

122

Figure C- 4. Florida ICW Transections for Counts of Reptiles of Special Concern

Abstract (if available)

Abstract The goal of this project is to develop an Ethington Transection Toolbox (ETT) to automate, increase the efficiency, and improve the efficacy of creating “Ethington Transections.” Ethington presents these hybrid charts/maps to visualize social change in space and time, along urban streets in his forthcoming book, Ghost Metropolis: A Global History of Los Angeles since 13,000. A new technique for visualizing the act of moving through the landscape over time, “Ethington Transections” are defined as a cross-sectional sample of data from polygons to simulate a single, directional line of transit. The objective of this thesis is to streamline the creation of transections resulting from the input of common polygon-distributed data, and to share such a tool so that others may benefit from increased efficiencies. The final result is a custom toolbox in Esri’s ArcGIS ModelBuilder of seven custom models with contextual help, written documentation and video walkthroughs. This series of models creates an editable map and graph layout and an organized geodatabase of intermediate outputs that can be reused for additional analyses or presentations. The ETT shortens the time to complete an “Ethington Transection” from 8 hours to slightly less than 1 hour. The previously tedious and time intensive task of creating transections was automated and made accessible to a wider range of researchers, facilitating new perspectives and interpretations of data. Therefore this toolbox should enhance the analytic skills of those looking to study how changes occur through space and time along any linear sample of data to simulate a transit in polygonal datasets.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Assessing the impact of a web-based GIS application to promote earthquake preparation on the University of Southern California University Park Campus

PDF

Finding environmental opportunities for early sea crossings: an agent-based model of Middle to Late Pleistocene Mediterranean coastal migration

PDF

Network accessibility and population change: historical analysis of transportation in Tennessee, 1830-2010

PDF

Eye.Earth Pro (Beta v1.0): application development and spatial financial analysis utilizing the PESTELM framework

PDF

Using pattern oriented modeling to design and validate spatial models: a case study in agent-based modeling

PDF

Peoples of Washington historical geographic information system: geocoding culture using archival standards

PDF

Precipitation triggered landslide risk assessment and relative risk modeling using cached and real-time data

PDF

Building a spatial database of biochar research and practice with Web-GIS

PDF

Geospatial web application development to access irrigation asset data: Veterans Affairs Palo Alto Health Care System

PDF

Developing an archaeological specific geodatabase to chronicle historical perspectives at Bethsaida, Israel

PDF

Visualizing historic space through the integration of geographic information science in secondary school curriculums: a comparison of static versus dynamic methods

PDF

Exploring urban change using historical maps: the industrialization of Long Island City (LIC), New York

PDF

Wrong way driving in San Antonio, Texas: a transportation route study using Network Analyst

PDF

Identification and analysis of future land-use conflict in Mecklenburg County, North Carolina

PDF

Cal ToxTrack: a full stack Web GIS for mapping pollution in California

PDF

Design and implementation of an enterprise spatial raster management system

PDF

A spatial analysis of beef production and its environmental and health impacts in Texas

PDF

Using GIS to identify potential dynamic marine protected areas: a case study using shortfin mako shark tagging data in New Zealand

PDF

Geospatial analysis of unintended casualties during combat training: Fort Drum, New York

PDF

A contributory Web-based application for documenting historic resources: French colonial era art deco architecture in Hanoi

Asset Metadata

Creator Bengoa, Anne Jeanene (AJ) (author)

Core Title Automating “Ethington Transections”: a new visualization tool

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 05/24/2016

Defense Date 09/11/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Ethington Transections,GIS,graphing,linear analysis,ModelBuilder,OAI-PMH Harvest,polygonal analysis,spatial analysis

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ethington, Philip (committee chair), Swift, Jennifer (committee member), Vos, Robert (committee member)

Creator Email a.jeanene.bengoa@gmail.com,bengoa@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-249438

Unique identifier UC11279673

Identifier etd-BengoaAnne-4418.pdf (filename),usctheses-c40-249438 (legacy record id)

Legacy Identifier etd-BengoaAnne-4418.pdf

Dmrecord 249438

Document Type Thesis

Format application/pdf (imt)

Rights Bengoa, Anne Jeanene (AJ)

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA