Katie Wholey

Intro. to GIS

December 16, 2009

Inner City Revitalization: Mapping and Predicting Gentrification in the Chicago

Background Information and Project Overview

At its very core, gentrification is a spatial issueoften defined as the transformation of neighborhoods from low value to high value. However, there is a certain level of ambiguity associated with the causes and effects of gentrification. Part of this elusiveness stems from the varying definitions of gentrification that exist within urban social science literature. According to The Encyclopedia of Housing: “Gentrification is the process by which central urban neighborhoods that have undergone disinvestment and economic decline experience a reversal, reinvestment, and in-migration of a relatively well-off, middle and upper middle-class population.” However, the Brookings Institute defines gentrification in a slightly different manner: “the process by which higher incomehouseholds displace lower income residents of a neighborhood, changing the essential characterand flavor of that neighborhood.” Thus, academic literature frames gentrification in a variety of ways; ranging from the broad characterization of neighborhoods as ‘gentrified’ based on the presence of urban revitalization in a particular area to more narrowly focusing on the physical upgrading of low-income neighborhoods and the resulting change in the demographic composition of the neighborhood, particularly in terms of racial and socioeconomic diversity.

As a result of this ongoing academic debate, this project analyzes several of the variables that academics assert as indicators of gentrification in order to predict which Census tracts will gentrify in the city of Chicago. In order to evaluate these effects, the simplest definition of gentrification will be used: high levels of renovation in low-income neighborhoods that leads to the in-migration of higher income residents.

Major Steps and Methodology

Initial Research

According to the academic literature, a wide range of variables characterize gentrification. Some variables are related to the changing demographics of the population in an area while others are related to the physical attributes of the area and the housing stock. Therefore, I had to settle on a definition as well as a logical means of measuring gentrification. For the purposes of this project, gentrification is defined as high levels of renovation in low-income neighborhoods the leads to in-migration of the middle class.

Based on the studies by Helms (Helms 2003) and Dewar, Basmajian, Alter and Law (Dewar, Basmajian, Alter and Law 2006), I concluded that the physical attributes of a neighborhood and its surrounding area play a large role in determining gentrification. Helms’ paper establishes that the characteristics of the building and its neighborhood play an important role in determining whether or not the building will be renovated.Along with Helms, Dewar, Basmajian, Alter and Law assert that early warning signs of gentrification lie almost entirely in the physical attributes of the neighborhood.

Furthermore, the GIS work of Gina Clemmer and Ashon Nesbitt were used to understand how GIS can be used to assess gentrification and also helped to define the most important indicators of gentrification. Clemmer finds that census tracts often display varying degrees of gentrification. Clemmer also begins to establish a technique by which researchers can concretely identify changing neighborhoods and the level of gentrification experienced by each of these neighborhoods (Clemmer 2000). Nesbitt (Nesbitt 200%) builds on this idea and technique by constructing a predictive model that establishes sixteen accepted indicators of gentrification, nine of which are used in this analysis. Thus, largely as a result of these four studies and additional literature (listed in the bibliography), I identified the indicators used in this study; notably, the characteristics often associated with gentrifiers include educational attainment, tenure,race,and median age while those associated with physical changes to a neighborhood focus on vacancy rates, size of housing units, age of housing stock, totally population of the neighborhood, and proximity to high income tracts.

Methodology

  1. I obtained GIS layers for the city of Chicago from ESRI and the City of Chicago GIS database. These layers included Census tracts for the city as well as SF1 data from the 2000 Census for the city of Chicago (by Census tract).
  2. I created an Excel spreadsheet from the building permit data for the year 2000 found on the website of the Chicago Metropolitan Agency for Planning. Once the data was in an acceptable format, I removed any data for non-residential buildings. Although the upgrading of commercial areas is a large part of gentrification, the analysis for this project is based on gentrification in relation to residential areas and the people that often gentrify in any particular area; thus, the building permit data was limited to residential buildings. I then added the table to GIS and geocoded the building permit data by address. Overall the geocoding was successful with 92% of the data matched.
  3. A spatial join was performed, joining the geocoded building permit data to the Census tracts in the city in order to aggregate the building permit data by tract.
  4. Census data from the year 2000 was also downloaded into Excel and then imported into ArcMap so the identified indicators of gentrification could be analyzed using GIS. This data was then joined to the Census tracts according to the TRACT field.
  5. With all of the data correctly defined and imported into ArcMap, I converted each of the layers to the same projection: NAD 1983 State Plane IllinoisEast FIPS 1201 in Feet.
  6. Then, I mapped each of the eight indicators of gentrification and GIS sorted the data for each indicator into quintiles.
  7. For each of the eight indicators, I added a field in the attribute table in order to assign a gentrification risk score to each Census tract. I then used “select by attribute” to select the bottom two quintiles, the middle quintile, and the two highest quintiles of each indicator. Next, I used the field calculator to assign a score between one and three to each of the selected groupsaccording to the gentrification risk of each indicator.
  8. I also used “select by location” to select the low income tracts directly adjacent to a high income tract, as characterized by those tracts that are at least 0.5 standard deviations above the median income for the city of Chicago. I created a new field for this attribute and assigned a score of 1 to each tract that touched a high income tract and a score of 0 to those that did not. Neighbors often play a large role in gentrification, so the presence of presumably wealthier, more educated, higher class individuals should impact a neighboring tract and cause it to gentrify more rapidly than a low income tract that is not directly adjacent to a high income tract.
  9. Finally, I used the field calculator to aggregate the scores of each Census tract across all indicators in order to obtain an overall risk of gentrification for each Census tract in the city of Chicago.
  10. However, because gentrification, by definition, can only occur in low income neighborhoods, I then used “select by attribute” to select out the low income Census tracts. I defined low income as those tracts that were 0.5 standard deviations or more below the median income for the city of Chicago. Then, I created a layer from the selected low income tracts and mapped the overall risk of gentrification across these tracts.
  11. Finally, I mapped the building permit data across these low income tracts in order to assess how accurately the indicators of gentrification predicted the actual renovations in each Census tract.
  12. Furthermore, I also used the spatial analyst to create a density surface map of the building permit data for these low income tracts in order to show how the impact of upgrades and renovations can extend beyond the Census tract that they take place in; a very important spatial aspect of gentrification is how surrounding areas can impact the physical upgrading of a neighborhood.

Reasoning for Including Each Variable.

-Building permit valueper household can be used as a means by which to assess the amount of physical upgrading that has taken place within a census tract.

-Educational attainment is a key indicator of gentrification because education is correlated with higher income; thus, gentrifiers are typically more educated than non-gentrifiers.

-Because there is often a high level of minorities in low income areas, race can be used as a means by which to identify tracts vulnerable to gentrification; tracts with a large white population are generally less likely to gentrify than those with a small white population.

-While median age per census tract is not considered to be one of the main indicators of gentrification, it is an important descriptive characteristic because older populations generally have the means by which to upgrade their housing and also are more likely to be owners instead of renters.

-Tenure plays an important role in gentrification. Because renters are not personally invested in their household unit, they generally do not place high importance on the physical upgrading of their households. Owners, on the other hand, place a much higher importance on the upkeep of their house and thus, areas with large amounts of renters would be expected to have lower levels of gentrification.

-Additionally, it is important to look at the relationship between vacant units and gentrification because neighborhoods with high vacancy rates have more available housing stock to be upgraded and thus, may be more likely to be upgraded because the unit can be easily obtained by gentrifiers.

-Smaller houses have more potential for upgrade; thus, median rooms is used as an indicator of gentrification.

-The age of the housing stock identifies housing units that are in greater need of physical upgrading and thus, more likely to be gentrified.

-Population determines the density of a neighborhood and density is often strongly correlated with neighborhoods that are in need of renovation because a lot of people living in a comparatively small area often signifies overcrowding and thus, inadequate housing arrangements.

Data Sources

The Census tract data layers came from the city of Chicago’s GIS database. Additionally, the building permit data was obtained from the Chicago Metropolitan Agency for Planning. Lastly, the SF1 Census data was supplied at the Census tract level by ESRI and the SF3 data was downloaded directly from the Census Bureau and then imported into ArcMap.

Potential Limitations, Difficulties, and Concluding Thoughts

The major challenge of this project was to create the building permit data set and get it to join to the Census tracts correctly. The Census tracts were identified differently in the GIS layer than they were in the building permit data. The building permit data listed Census tracts by the simplest from of the tract number (i.e. 101) while the Census tract layer listed the tracts by their six digit tract number (i.e. 010100). Thus, I had to convert the tract numbers in the building permit data to the six digit form. However, Excel automatically removes leading zeroes on numbers; thus, I had to work around this complication. Even when I was able to get Excel to keep the leading zeroes, ArcMap would remove them when I added the data table. Eventually I discovered that I could convert the number to text and both Excel and ArcMap would leave the leading zeroes so that I could use the TRACT field to join the data to the Census Tract layer. Another difficulty I had is evidenced in my maps. Because I took my GIS layers from the city of Chicago they do not match up well with the Tiger hydrography layer. I struggled with the discrepancies between the two layers and eventually decided to leave off the context around this city of Chicago for my final maps; however, I am well aware that there should be context in terms of surrounding towns and also hydgrography.

Additionally, because of the nature of the issue of gentrification, there are several potential issues with my analysis. The argument could be made that I should have removed high income tracts from the beginning of my analysis instead of at the end of my analysis. However, I decided that it should be removed at the end because of the effect that surrounding neighborhoods can have on their neighbors. Therefore, I thought it was important to look at the overall spatial distribution of each of the gentrification indicators that were mapped. Additionally, each of the variable used were given the same weight for their effect of the overall gentrification risk of each tract; in reality, this is probably not the case. Furthermore, gentrification is largely based on the idea of changing neighborhoods in that low income neighborhoods are upgraded and then turn into middle class neighborhoods. Thus, my study would have been more effective if I analyzed the data across several time periods. However, I did not have the building permit data for this type of analysis. Lastly, the data is fairly old so the analysis does not pose much merit at the present time with new Census data being made available within the next year.

Perhaps the most important issue of this analysis is based upon the nature of gentrification itself. The argument can be made for several variables that I have left out of my analysis, including proximity to public transportation, proximity to the central business district, upgrading of commercial areas located near each Census tract, percentage of subsidized housing units, etc. Thus, while the analysis is an interesting exploration of the nature of gentrification, it cannot be taken as an exact prediction of gentrifying tracts in Chicago as there are many limitations to this sort of analysis. Furthermore, my lack of familiarity with the city limits my analysis because I cannot verify my results.

References

Brookings Institute. brookings.edu (accessed November 5, 2009).

Brueckner, Jan K., and Stuart S. Rosenthal. 2005. Gentrification and Neighborhood Housing Cycles: Will America’s Future Downtowns be Rich? CESIFO Working Paper No. 1579, (10) (October 2005): 1-45.

Census Bureau. 2000 Census.

Chicago Metropolitan Agency of Planning.

Clemmer, Gina. 2000. Quantitative and Spatial Analysis Techniques for Analyzing Gentrification Patterns, Case Study: Portland, Oregon. PortlandStateUniversity Independent Research Project (September 2000): 1-16.

Dewar, Margaret, Carlton Basmajian, Rebecca Alter, and Betty Law. 2006. Early Detection of Changing Neighborhood Conditions. University of Michigan Urban and Regional Research Collaborative: 1-23.

Freeman, Lance. 2009. Neighbourhood Diversity, Metropolitan Segregation and Gentrification: What are the Links in the US? Urban Studies 46, (10) (September 2009): 2079-101.

———. 2005. Displacement or Succession? Residential Mobility in Gentrifying Neighborhoods. Urban Affairs Review 40, (4) (March 2005): 463-91.

———, and Frank Braconi. 2004. Gentrification and Displacement: New York City in the 1990s. Journal of the American Planning Association 70, (1) (Winter 2004): 39-52

Guerrieri, Veronica, Daniel Hartley, and Erik Hurst. 2009. Endogenous Gentrification and Housing Price Dynamics. University of Chicago, BoothSchool of Business (October 16, 2009): 1-58.

Helms, Andrew C. 2003. Understanding Gentrification: An Empirical Analysis of the Determinants of Urban Housing Renovation. Journal of Urban Economics 54, (3) (November 2003): 474-98.

Jennings, James. 2004. Urban Planning, Community Participation, and the Roxbury Master Plan in Boston. Annals of the AmericanAcademy of Political and Social Science 594: 12-33.

Kennedy, Maureen, and Paul Leonard. 2001. Dealing with Neighborhood Change: A Primer on Gentrification and Policy Choices. The BrookingsInstitutionCenter on Urban and Metropolitan Policy and PolicyLink (April 2001): 1-70.

Kolko, Jed. 2007. The Determinants of Gentrification. Public Policy Institute of California (December 2007): 1-29.

Lees, Loretta. 2009. Urban Renaissance in an Urban Recession: The End of Gentrification? Commentary. Environment and Planning A 41, (7) (July 2009): 1529-33.

———. 2008. Gentrification and Social Mixing: Towards an Inclusive Urban Renaissance? Urban Studies 45, (12) (November 2008): 2449-70.

London, Bruce, Barrett A. Lee, and S. Gregory Lipton. 1986. The Determinants of Gentrification in the United States: A City-Level Analysis. Urban Affairs Review 21, (3) (March 1): 369-87.

Martin, Richard W. 2009. Gentrification in American Cities. University of Georgia (October 2009): 1-49.

Nesbitt, Ashon J. 2005. A Model of Gentrification: Monitoring Community Change in Selected Neighborhoods of St. Petersburg, Florida Using the Analytic Hierarchy Process. University of Florida Urban and Regional Planning GraduateSchool Thesis: 1-72.

Van Vliet, Willem, ed. The Encyclopedia of Housing. Thousand Oaks, CA: Sage Publications, 1998.

Vigdor, Jacob L., Douglas S. Massey, and Alice M. Rivlin. 2002. Does Gentrification Harm the Poor? [With Comments]. Brookings-Wharton Papers on Urban Affairs: 133-82.