Quantifying Correlation of Food Access and Mortality via Regularized Linear Models & Random Forests
Food Access & Security in the United States is commonly characterized as a seemingly limited problem only impacting isolated rural communities. However, as of 2010, there are more than 140 million people living in urban and suburban communities that are more than 0.5 mile (urban) / 10 miles (rural) from a supermarket. Additionally, there are more than 60 million people in suburban and rural areas that are more than 1 mile (suburban) / 20 miles (rural) from a supermarket. According to a report submitted Congress, the USDA has acknowledged that “limited access to nutritious food and relatively easier access to less nutritious food may be linked to poor diets and, ultimately, to obesity and diet-related diseases.” The issue is inflated when residents are low income and lack access to vehicles.
In this study, I examine the correlation between mortality rates and food access. By isolating the coefficients of determination, beta coefficients and p-values from linear and poisson regressions across several suburban and rural demographics, we can see how susceptible communities are to disease and increased mortality rates.