Income Disparity & Public Transport in the Atlanta Area

Objective: Visualize correlations between wealth distribution and access to public transportation, in order to:

1. Better understand the connections between Atlanta's history and its infrastructure
2. To identify potential inequalities and propose policy solutions
3. Draw conclusions about cultural attitudes toward public transport in different regions

Hypothesis: Access to public transportation will be negatively correlated with extremes in wealth. There will be a disparity in access among lower income areas, and a rejection of public transport in higher income areas.

Data Acquisition:

• Census tracts for the counties of Clayton, Cobb, DeKalb, Fulton, and Gwinnett from the U.S. Census Bureau's official website.
• Tables for 2022 household income statistics by census tract for the same counties, from the same website.
• "Transit routes" found on ESRI, uploaded by Georgia Association of Regional Commissions (originally was going to create my own using various tables of bus and train stops, but this data was more convenient and accurate)

Methodology:

• First, census tracts as foundation
• Then, joined data from household income tables to census tract attribute tables (some done manually due to technical issues)
• Created 1 mile radius buffer from transit routes shapefile
• Clip function to clear buffer that exists outside of the 5 counties
• Clip again to create 5 buffer layers, 1 for each county
• Beige-to-green color scale for tracts to represent lower and higher amounts of money (green = money), and to serve as a neutral backdrop for buffer layer
• Blue buffer layer for perfect balance of standing out without obstructing readability for rest of map
• Animations you see here by creating a .gif of 2 layouts fading into one another for each county: 1 without buffer layer, and 1 with buffer layer
• Cleaned up map by removing unintentional polygons and tracts with $0 household incomes (prisons, airports, etc)
• NAD 83

Results: Now, let's take a look at the data to further analyze the relationships between wealth and public transport. (Note: "entirely accessible" = entire tract is within 1 mile of a public transit route, "partially accessible" = at least part of the tract is within 1 mile of a public transit route, "inaccessible" = no part of the tract is within 1 mile of a public transit route)

• The smallest of the 5 counties, and features the smallest range in income
• It has the lowest overall median income of the 5 by a wide margin
• Notice the correlation between public transit routes (blue) and lower median household incomes?
• Every tract with a median household income of $60,000 or less is at least partially within 1 mile of a public transit route

• Perhaps most prominent correlation of all 5 counties
• The most traditionally suburban of the counties
• East Cobb and West Cobb have significantly higher median incomes than the city of Marietta, and also are less accessible by public transport
• Of 16 tracts with a median household income of $170,000 or higher, 12 are inaccessible by public transport
• Again, every tract with a median household income of $60,000 or less is at least partially accessible

• Can be seen as an inverse of Cobb County, where the highest income tracts are actually the most likely to be accessible by public transit
• Correlation between low incomes and public transport access is weak toward the city center, but strong toward the eastern suburban parts of the county
• Once more, every tract with a median household income of $60,000 or less is at least partially accessible

• Notable divide in transport access between wealthy urban areas in central Fulton (Buckhead, Sandy Springs) and wealthy suburban areas in northern Fulton (Milton, Alpharetta, Johns Creek)
• These used to be separate counties (Milton County was annexed by Fulton County in 1932)
• The lower income tracts in the city of Atlanta are among the poorest in the state — but notice that they are entirely covered by blue here?
• Another county where every tract with a median household income of $60,000 or less is at least partially within 1 mile of a public transit route

• Similarly to Cobb County, low income tracts are pretty prominently correlated to highways and bus routes
• Finally: 2 tracts with a median household income of $60,000 or less are inaccessible, the only county where this happened
• Still, the other 42 are at least partially accessible

Some interesting statistics to take note of:

• Of 239 census tracts with a median household income of $60,000 or lower, 237 have at least partial access (as defined in this project) to public transportation, for a rate of 99.16%.
• Of 102 census tracts with a median household income of $170,000 or higher, 67 have at least partial access to public transportation for a rate of 65.69%.
• DeKalb County is an interesting outlier, with all 34 of its $170,000+ tracts falling entirely
• In every county except for Fulton, proximity to interstates is strongly correlated to lower incomes.

Some broad conclusions I've arrived at over the course of analyzing this data:

• My hypothesis was only partially correct. Access to public transportation is pretty ubiquitous in lower income areas, but, as I predicted, it is noticeably lacking in some specific higher income areas
• In contrast to MARTA, Cobb County and Gwinnett County's independent transit systems seem to be more focused on providing necessary transportation options for low income communities, rather than providing convenient access to all parts of their county. Which is the correct focus?
• While wealthy urban areas embrace public transport, wealthy suburban areas seem to reject it — "spreading of poverty" narrative?
• DeKalb County: highly wealthy and extremely poor tracts all being easily accessible by public transport — Does it serve as an example that the previous notion is misguided?

Limitations:

• Assumes household income is a valid indicator of wealth
• Biggest one: I should have used way more census data, not just household income! Lots of room for interesting analysis with other economic data, demographic data, etc.

Thank you!