A Spatial Dependency and Causality Analysis of Crime

This article examines the spatial variability in the distribution of crime in the City of Savannah, GA, where a total of 12,458 crimes were reported in 2000. All crimes were geocoded and the same numbers of control locations were generated under the assumption that the spatial distribution of crime is a realization of an inhomogeneous Poisson process (spatially random). The control locations served as a basis of comparison from which individual crime data were modeled as a function of proximity to alcoholserving establishments and socio-economic data, including, income, race/ethnic segregation and education. As expected, crimes locations were highly clustered and among a host of covariates proximity to alcohol-serving establishments, household income and ethnic diversity emerged important controls of crime incidences in the study area.