Health Data: This was obtained from the 500 cities study that was conducted by the Center for Disease Control (CDC) for a number of adverse health outcomes. Obtained by census tract, the data captures the prevalence of adverse health outcomes for a number of diseases in each tract. For example, prevalence of coronary heart diseases in census tract A might be lower than that in census tract B.

Infrastructure Data: This data set was broken down further into two categories, green space and roadways. The green space data set consists mainly of green amenities such as tree canopy and parks, measured respectively as the area percentage in each census tract area and green space access. The metric for roadways here, was taken as the degree of connectivity of roads in each census tract to serve as a proxy for traffic.

Demographic Data: Taken from the US census bureau, demographic variables such as the proportion of ethnic makeup (% African American) and percentage of residents in particular age categories (i.e. over 65 years) by census tract was gathered from the 2010 data set.

Air quality Data: Modeled at fine spatial scales of 250m with chemical transport and dispersion models, the data was later aggregated to the census tract level. The modeled data was available for 2010 for carbon monoxide (CO), nitrogen oxides (NOx) and fine particulate matter (PM2.5)

Plots: Plots of the data set can be found in the publication listed in the ‘associated publications’ section.

The data was available within 2006 to 2016, which spans the time scope of the study. The air quality and demographic data were obtained for 2010 and the health data became available in 2015, although it was based on historical data and statistical estimations.