Outcome Variable: crime rate (2010 for predictive and 2016 for implementation)
Predictor Variables: socioeconomic status and urban planning features (street connectivity, school counts, land cover, and pollution sites)
Top 3 Important Variables for Violent Crime: proportion of people non-Hispanic Black, Disadvantage 1, and proportion of families with Income less than 15K.
Top 3 Important Variables for Property Crime: Disadvantage1, Proportion of female-headed families with kids, and Disadvantage2.
Results
Violent Property
Final prediction for violent crime rate in 2016 with XG Boost has a RMSE of 0.593.
Final prediction for property crime rate in 2016 with Random Forest has a RMSE of 0.459.