Actionable Insights on Philadelphia Crime Hot-Spots: Clustering and Statistical Analysis to Inform Future Crime Legislation
share
Philadelphia's problem with high crime rates continues to be exacerbated as Philadelphia's residents, community leaders, and law enforcement officials struggle to address the root causes of the problem and make the city safer for all. In this work, we deeply understand crime in Philadelphia and offer novel insights for crime mitigation within the city. Open source crime data from 2012-2022 was obtained from OpenDataPhilly. Density-Based Spatial Clustering of Applications with Noise (DBSCAN) was used to cluster geographic locations of crimes. Clustering of crimes within each of 21 police districts was performed, and temporal changes in cluster distributions were analyzed to develop a Non-Systemic Index (NSI). Home Owners' Loan Corporation (HOLC) grades were tested for associations with clusters in police districts labeled `systemic.' Crimes within each district were highly clusterable, according to Hopkins' Mean Statistics. NSI proved to be a good measure of differentiating systemic (< 0.06) and non-systemic (≥ 0.06) districts. Two systemic districts, 19 and 25, were found to be significantly correlated with HOLC grade (p =2.02 × 10^-19, p =1.52 × 10^-13). Philadelphia crime data shows a high level of heterogeneity between districts. Classification of districts with NSI allows for targeted crime mitigation strategies. Policymakers can interpret this work as a guide to interventions.
READ FULL TEXT