Newest Batch of GIS-Related Theses & Dissertations

Have not posted a new batch of dissertations or theses since the summer, so here are the latest dissertations & theses that have caught my eye. Some of these are really quite fantastic.

Most of the following links include a free 24-page preview of the dissertation.

Previous lists include 11/05, 01/06, 01/06, 04/06, 07/06

• Decolonizing geographic information systems, by Eades, Gwilym Lucas, MA, Carleton University, 2006, 107 pages.
• "This thesis consists of four parts, each of which contributes to answering the primary research question: to what extent are Geographic Information Systems (GIS) a colonial technology?...By way of conclusion, a process of decolonization is theorized, focusing on the lands of Indigenous peoples and on the practices of GIS 'experts' situated within Indigenous communities. "
• The capitalization of exogenous features into the sale price of single family homes, by Ehlen, Victoria A., MA, State University of New York at Binghamton, 2006, 105 pages.
• "This thesis examines the role of exogenous housing features in predicting the sale price of homes, using data acquired for the Greater Binghamton area. With GIS and statistical methodology, a regression equation was generated to predict sale price based on endogenous and exogenous variables."
• The built environment and walking: An examination of school transport mode, by Dreyling, Erin Kyle, PhD, The Johns Hopkins University, 2006, 241 pages.
• "It was concluded that research on walking can benefit from greater use of these tools in several ways. First, GIS can facilitate environmental data collection by aiding in the identification of locations to examine in a study area. Second, GIS and spatial statistics can help investigators generate hypotheses about the environmental determinants of walking that may go unnoticed in nonspatial analysis. Third, spatial statistics can improve regression modeling by ensuring models account for spatial dependence."
• Assessing the impacts of the 2004 tsunami on mangroves using GIS and remote sensing techniques: A case study of Phang Nga, Thailand, by Sirikulchayanon, Poonthip, MS, Southern Illinois University at Carbondale, 2006, 102 pages.
• "The study focuses on the spatial association between mangrove forests and the damage. Using GIS and remote sensing techniques, we compile the data from various sources and detect the land cover changes to determine the level of damage in regions with and without mangroves. The proposed novel approach integrates GIS proximity analysis and change detection with remote sensing techniques to creatively delineated buffer zones from the coastline into four homogenous sub-regions."
  
• Quantification and delineation of the nonattainment boundary for fine particulate matter: Using geographic information system (GIS), remote sensing data, and in situ monitoring, by Rush, Alan C., PhD, George Mason University, 2006, 183 pages.
• "This dissertation describes the work of developing a tool for use in delineating the boundaries of nonattainment areas as defined by the Clean Air Act. The weakness of the ambient monitoring system is discussed, and the need for an advanced method for pollutant assessment is concluded."
• Student discipline referral rates and neighborhood crime statistics, by Cash, Tamara G., PhD, The University of Kansas, 2006, 133 pages.
• "Students' gender, grade level, and ethnicity variables were analyzed, and GIS software was used to geocode or map the students' home addresses and the crime locations. Results found a low positive correlation between the frequency and severity of student infractions and the frequency and severity of crime incidents per cell location. An inverse relationship was found between student infraction frequency and severity and the students' proximity to high-crime areas."
• A battlefield preservation plan for Champion Hill (Mississippi), by Dean, Leslie Alan, MLA, Mississippi State University, 2006, 272 pages.
• "This plan creates a long-range vision for preservation of Champion Hill while focusing on specific and short-term measures. This study employed GIS technology as a tool for gathering, recording, and analyzing important site data based on McHarg (1969), Hopkins (1977), and Drummond (1998). This study concludes with eleven recommendations for the preservation of the Champion Hill Battlefield area. "
• Urban growth and segregation in the Roanoke, Virginia, metropolis: The effects of low-density development on low-income populations and racial minorities, by Etienne, Freed G. (Mike), PhD, Virginia Comonwealth University, 2006, 173 pages.
• "...[W]e must understand the social consequences to urban life, relative to concentration of poverty and racial minorities in central cities. Toward that end, this study uses the statistical techniques called Social Area Analysis and Factorial Ecology to examine and describe the social-spatial patterns of the Roanoke, Virginia, metropolis, focusing on poverty and race. "
• Online public participation GIS. Shaping the scales and spaces of comprehensive planning: A case study on the use of geographic information systems on the Web to support comprehensive planning in the town of Amherst, New York, by Howard, Daniel C., PhD, State University of New York at Buffalo, 2006, 341 pages.
• "Comprehensive planning is an opportunity for communities to collectively envision their future through a communicative participatory endeavor that can establish a foundation for lasting civic involvement. Comprehensive planning involves a process that guides the future designation of geographic spaces through social action; this process can simultaneously be empowering for some, and marginalizing for other stakeholders."

Oh yes, and lest we forget, everyone have a frightful Halloween, OK?

Flipping Properties: GIS Can Show You Where

Developed a new project for my graduate real estate students to chew on this semester. Working in teams of 2 to 4, the students need to identify the top two zip codes in the D/FW Metroplex to invest in residential properties in order to flip them over.

<-- Map showing the location of every residential sale in 2004.

I organized the course around three projects. The first project was focused on data acquisition and creation. The second project, this property flipping project, is focused on site selection and analysis. The third project will be a full market research project for a proposed new hotel development in the entertainment district of Arlington, TX.

So, what would make an area more suitable than another to invest in the residential real estate market? Well, these students must consider the following three variables:

1. Days on the market. This is the average number of days properties take to sell once the property is listed.
2. Property value change. This explores how much the values of the properties are increasing or decreasing.
3. Population change. This explores how much the population is increasing or decreasing.

The students then need to add a minimum of three additional variables and use ESRI's Spatial Analyst extension to identify the top two zip codes.

So, how can GIS help?

• First we had to geocode the local MLS data (which created the map above).
• Playing with MLS data in GIS is so much fun. The students were provided with a table containing every property sale in 2004 in the metroplex. There are over 45,000 sales during this year. For each sale, the address, days on the market, list price, sale price, interest rate, mortgage amount, and many, many more variables beyond what is normally in a standard appraisal district file.
• Second we had to aggregate the average number of days on the market and the average selling price during the year to the block group level.
• These attributes were aggregated so they could be correlated and analyzed with various demographic data on the block group level.
• Third the teams are all going in various directions with the project.
• Some are manipulating the datasets to explore fluctuations over the year. Some are acquiring additional sources to determine the number of vacant units and the number of nearby rental opportunities.
• Fourth the teams will have to use suitability analysis to locate the best spots that match their criteria.

Project is due in two weeks and I am eagerly looking forward to see what they come up with.

I have been pondering the idea of tailoring this project into a 2-hour workshop because of the attractiveness of performing this type of research. However, we can not distribute the MLS data and I would hate to hold a workshop and then tell everyone that they could not go out and repeat what they just learned how to do. Ah well...I'm still trying to think of some way to do it.

Workshop: Low Income Housing, Single Mothers, and GIS

Bibliography of Low-Income Housing Data (annotated, with a focus on Texas)
Entire Workshop Materials (Full CD)

Held a workshop last week focusing on low-income housing. The purpose of the workshop was to increase awareness of the amazing datasets (GIS and tabular) available on the HUD User website, as well as numerous others. As is evident by perusing through the annotated bibliography, I compiled a fairly comprehensive list of resources. Data source agencies include HUD, Texas Department of Housing & Community Affairs, United Way, Census 2000/1990, Texas State Data Center, and AGS demographics.

The workshop was focused around a fictitious foundation offering grants to developers of low-income multifamily housing units targeting single mothers. Using single mothers as the target population helped to increase the attraction of the workshop and gave us a chance to play around with some interesting and relevant demographic data. We focused on the 6-county area of Collin, Denton, Tarrant, Dallas, Johnson, and Ellis.

Had 16 folks show up, which is pretty phenomenal considering the scholarly and non-popular focus of the workshop. A rep from the Dallas office of the Fannie Mae Foundation (not a fictitious foundation) attended and seemed to have a good time. A whole group of social work students attended, as well as real estate (which was a pleasant surprise). Also was a handful of social work faculty and a nursing professor as well.

A CD image of the workshop materials can be downloaded here. ArcGIS 9 is required.

Census 2000, Voting Precincts, & Parcel Geocoding

Hoo boy do I have a lot of catching up to do, but I'll start with this...

Professor emailed a table to me containing app. 15, 000 addresses within Arlington, TX, and wanted to know:

1. Voting precinct the address falls within
2. Voting precinct centroid the address is closest to
3. Distance between the address and the closest precinct centroid
4. Various Census 2000 demographics for each precinct
   

This was a fun project that took no more than a couple hours. Here is what I did.

1. Geocoded the addresses to the city of Arlington parcel boundaries. As discussed here, the increased accuracy is worth the extra effort than relying on street geocoding. StreetMap USA makes it super-easy, but the errors compound themselves. Match rate was above 95%. Very excellent. The Arlington parcel boundaries can be pulled down here.
   
2. In ArcMap, spatially joined the voting precinct boundaries to the geocoded addresses. This gave me the precinct number for each address. The redistricting boundaries can be pulled down from the Texas Legislative Council's FTP server
3. In ArcMap, used the Feature to Point tool to generate the precinct centroids.
   
4. In ArcMap, used the Near tool to calculate the closest (nearest) precinct centroid to each geocoded address. This also provides the distance, in feet, as I projected the data in State Plane (feet).
5. Pulled down the necessary Census 2000 stats from Factfinder and joined it to the block groups.
6. In ArcMap, spatially joined the block groups to the voting precincts. This aggregates the demographics on the precinct level based on where the block groups intersect the precinct boundaries.

That's it.

During the first half of this semester, I am receiving as many political science-related data requests as city planning and earth sciences, which offer degrees and/or certificates in GIS.