This visualization explores redlining and discriminatory federal New Deal housing policy’s lasting impact on homeownership, home equity, and wealth building opportunities across New York City.
In this data visualization thesis project, I explore the lasting impact of redlining and discriminatory federal New Deal housing policy on homeownership, home equity, and wealth building opportunities across New York City. I connect 1938 government redlining maps of the five New York City boroughs and lower Westchester county with longitudinal census data from 1940 to 2010, tracking changes in racial composition and real estate markets within the neighborhoods included within these maps. I identify specific case study comparisons between neighborhoods with similar population demographics or homeownership markets in 1940, connect these communities to the redlining gradings they received at that time, and follow their changing and diverging housing markets through 2010. My preliminary qualitative analysis of these case studies suggests that property values and homeownership rates were depressed in many neighborhoods across New York City that received poor redlining grades or did not retain a majority white population in the decades after these official government security maps were drawn. Finally, I open up the map for exploration, allowing users to study the redlining map of New York in more detail or investigate potential threads of housing discrimination themselves.
Data Sources and Methods
Redlining zone boundaries, gradings, and area descriptions (including apprasier quotes) are from the Mapping Inequality project. Data was downloaded in GeoJSON format for Manhttan, Brooklyn, Queens, the Bronx, Staten Island, and lower Westchester county.
Census data and tract boundaries for each decennial census from 1940 to 2010 are from the IPUMS NGHIS database. Included in census data by tract were tabular data tables used to estimate percent non‑white, percent homeownership, and median home value.
Census data by tract were joined to redlining zones using geospatial geometric analysis made possible with the Shapely Python library. Census data estimates were weighted by the proportional size of a tract's geographic area of intersection with a redlining zone. Data was averaged across multiple redlining zones for neighboorhoods in case studies.
Geographic boundaries of New York City boroughs are from dwillis' nyc-maps GitHub repository
Scrollytelling made possible with Scrollama.
Building image comparisons made possible with JuxtaposeJS.