This study investigates the correlation between demographic features and social distancing adherence (SoDA) across US counties. Using mobile phone data and population statistics from 3054 counties, the researchers identified key demographic features correlated with SoDA. A multivariable bagging regression algorithm was developed to predict SoDA, achieving 90.8% accuracy. The study highlights the impact of economic, health, and political factors on SoDA and proposes the prediction model as a valuable tool for informing health policy and interventions.

Myles Ingram, Ashley Zahabian, Chin Hur