Every spring, the University of Wisconsin Population Health Institute releases a massive Excel file of health data for every county in America, which they publish on CountyHealthRankings.org. The data has everything from traditional risk factors like smoking and obesity to information on health insurance coverage, access to healthy foods, exercise, and mental health. This sort of dataset is a journalist’s dream, because it draws together dozens of disparate sources and subjects each one to the scrutiny of statistical experts.

Limitations

A lot of health data lags by several years. The CDC’s Behavioral Risk Factor Surveillance System, for example, trails by about three years, so the 2015 version of CountyHealthRankings contains smoking data from 2012.

Since the majority of the sources are based on samples, there can also be very high margins of error for low-population counties.