Measuring macro and micro geospatial changes in New York City’s motor vehicle thefts

Abstract Objective To test the capability of a novel spatial analysis to track macro and micro level changes in where crime occurs after policy change. Methods We implement a pre-processing method for spatial data called “Distance of Distances”, to detect geospatial clusters of crime within a radius around a central point. Using COVID-19 “shelter in place” restrictions in New York City, we examine before-and after changes in the spatial distribution and clustering of MVT on both a macro (citywide) level and a micro-level surrounding acute care hospitals. Findings MVT gradually clustered more densely around the city center, and in outer regions, while cluster density declined at intermediate distances between the city center and the outer regions. Comparatively, MVT clusters quickly increased in distance from acute care hospitals, demonstrating that changes in MVT were neither temporally nor spatially uniform. Conclusion The analytical procedure of combining micro and macro effects yielded broader insights than may be found at only one level. This approach was useful for measuring spatial changes in offending across time, broadening the discussion on displacement and diffusion. The results offer a useful validation method for place-based analyses without the need to operationalize buffer regions.