A public health approach to violence prevention involves the empirical identification of groups and communities at the highest risk for violence to inform targeted interventions. We demonstrate the utility of complete incident-level crime data toward this end. Data for 32,056 unique incidents involving homicide, aggravated assault, and robbery were extracted from the 2013 Michigan Incident Crime Reporting system, a statewide National Incident-Based Reporting System (NIBRS) data system. Differential victimization rates were calculated across demographic subgroups and jurisdictions to identify patterns in risk. Two-stage least squares regression models were estimated to examine correlates of variation in excess risk. Analyses identified young Black males and females at relatively high risk for violent victimization, and that this risk was amplified within cities with disproportionately high crime rates. Multivariate models suggested concentrated disadvantage as the most stable correlate of variation in excess risk across Michigan cities and towns. The results highlight the importance of expanding NIBRS adoption and the deployment of focused interventions involving both short-term enforcement and long-term social reinvestment.
Big data analytics and Mining for effective Visualization and trends forecasting of Crime data
