AirHealthLink is a project dedicated to studying the link between economic condition and air pollution using publically-available levels of common pollutants (PM2.5). It integrates environmental data from the EPA Air Quality System (AQS) and economic data from the U.S. Census Bureau to quantify disparities in air pollution exposure.
Using 615 counties with complete data:
A consistent negative association was observed between PM2.5 and income, showing an estimated decrease of approximately $1,350 in median household income per 1 μg/m³ increase in PM2.5 (p < 0.001).
Lower-income counties experience approximately 0.8 μg/m³ higher PM2.5 concentrations compared to higher-income counties.
These results provide quantitative evidence of economic disparities in air pollution exposure across U.S. counties.
Data is taken from all available counties on the AQS API from January 1, 2022 to December 31, 2022.
County-level PM2.5 data was aggregated to daily summaries:
Annual: mean, median, standard deviation, 90th percentile.
Monthly averages.
Seasonal averages: DJF, MAM, JJA, SON.
Exceedance counts for thresholds of 12 μg/m³, 25 μg/m³, and 35 μg/m³.
AirHealthLink provides automated statistical workflows comprising of:
Correlation and Regression Analysis:
Pearson correlation,
Spearman rank correlation,
Ordinary Least Squares (OLS),
Robust regression (Huber),
Median quantile regression
Income Decile Evaluation:
ANOVA,
Kruskal–Wallis test,
Spearman rank trend test,
Monthly and seasonal regressions
Extreme Quintile Comparison:
Welch’s t-test,
Kolmogorov–Smirnov test,
Bootstrap confidence intervals
- Create a local clone of the repository.
- Sign up for an AQS API key and a U.S. Census API key and put them in src/private_keys.py, alongside the e-mail used to sign up for them.
- Run /src/main.py.
