Predicting Future Shooting Crime Locations Using Principles of Data Analytics (SHOPS)

Varlioglu, M. Said; Ozer, Murat

Shooting crime is a serious public health problem in the US. The analysis of any historical crime data reveals that crime is non-randomly distributed in time and space. Based on this notion, hot spots policing has gained its momentum to effectively predict future crime locations. Recent studies; however, pointed out that traditional hot spots policing occasionally predict rare crimes such as homicides and shootings due to their less frequent recurring counts in a given place and time (specifically for shorter time periods such as weeks and months). Given this context, we developed a new shooting prediction system (SHOPS) to explore whether recent dynamic/mobility activity patterns of known violent individuals increase the prediction of short-term fatal and non-fatal shootings compared to the traditional hot spots policing. Findings suggest that SHOPS predicts fatal and non-fatal shooting locations more precisely by identifying fewer hotspot locations. Policy implications of the study were discussed in the conclusion section.

创建者

证书

学科