To analyze the relationship between amenities and gentrification, a map displaying gentrification risk was created by combining real estate and demographic factors. Two housing-related factors (home values and rental costs) and four social factors (educational attainment, poverty level, Black population, and white population) were layered into a gentrification index, mapped at the census tract level. Amenity factors were then aggregated into subcategories, creating four potential explanatory variables: hospital community facilities, food, colleges/universities, and green spaces.

Before performing regressions, a Moran’s I analysis was performed on the gentrification index map to understand the spatial clustering or dispersal patterns present. The Moran’s I result suggested very slight clustering. An exploratory regression was then performed to identify potential variables for the final geographically-weighted regression. Though the exploratory regression did not identify any variables or variable combinations with a meaningful R-squared value, a geographically-weighted regression was performed using the variable combination with the highest value (R-squared = 0.13) to identify whether or not a meaningful relationship would be found when spatial autocorrelation was accounted for. The resulting geographically-weighted analysis also did not produce a meaningful R-squared value, suggesting the absence of relationship between our chosen variables and the gentrification map.

Despite not finding a significant relationship between measures of gentrification and the presence of amenities, our analysis did create a fairly reliable index for assessing likelihood of gentrification at the census tract level in the greater Atlanta area. This index could be used for future regression analysis to examine other factors with more data present and could be expanded upon to create an even more complex gentrification index. However, the results also highlight the complexities of assessing and understanding gentrification quantitatively, and thus our recommendation would be that future expansions on gentrification mapping and analysis include an aspect of qualitative feedback or use of community knowledge.