Mapping Gang Territory and Emerging Hotspots of Shootings
Kasturi, Anusha Sharma
MetadataShow full item record
Mapping is widely used as a tool for understanding the pattern of crimes and guiding intelligence led policing. This study examines gang member databases and recorded shootings over the period 2006-2015. Gang territories were mapped as polygons following a three step methodology which included kernel density estimation. Three sources of gang member locations were used (residential, arrest and incident locations). We observed that gang territories remained relatively constant throughout the study period and they overlap considerably. An early warning system for shootings was developed using Kulldorff’s spatial scan statistic (SaTScan v9.4.4). High Rate Proportion vs Total Incident Proportion plots were used to identify the SaTScan parameter combinations with best prediction accuracy. Clusters with no overlap, using 6 months data and a 1500m (1mile) buffer had the highest overall prediction accuracy of 75%. The detection of emerging clusters of shootings may be useful for designing targeted interventions to prevent gun violence.