Poster #244 - Understanding the Impact of Spatial Relationship Definitions on Crime Clustering Analysis: Implications for Urban Crime Patterns

• In Event: ASC Poster Session I

Abstract

Crime clustering analysis is a fundamental tool in understanding urban crime patterns and informing policing strategies. However, the choice of spatial relationship definitions—whether adjacency-based, distance-based, or network-based—can significantly influence analytical outcomes, potentially altering interpretations of crime hotspots and spatial dependencies. This study evaluates the impact of varying spatial relationship definitions on crime clustering metrics, specifically Moran’s I and Getis-Ord Gi*, to assess the sensitivity of crime analysis outcomes to these methodological choices. Using a crime dataset from a major urban center, multiple spatial configurations are applied to examine how clustering patterns shift under different spatial relationship definitions. Results indicate that the selection of spatial relationships can dramatically affect the identification and significance of crime clusters, with implications for resource allocation, crime prevention strategies, and policy-making. By highlighting the variability introduced by different spatial structures, this research underscores the need for methodological transparency and careful selection of spatial parameters in crime analysis. These findings contribute to improving crime mapping accuracy, advancing urban crime research, and enhancing evidence-based policing strategies.

Authors

• Hailey Nicole Richardson, University of Alabama, Laboratory for Location Science

• Kevin M. Curtin, University of Alabama, Laboratory for Location Science