Spatial Relationship and Colocation of Crimes in Jefferson County, Kentucky

Abstract

Crime is a major social issue in U.S. cities. Understanding the spatial pattern of crimes is imperative for developing informed policy and effective policing strategies that will reduce crime in local communities. This article examines the spatial distribution of major crimes in Jefferson County, Kentucky. Using crime data spanning the period between 2009 and 2011, this study looked into the spatial relationships and colocation patterns among different types of crime across the county. Specifically we tried to uncover which crime types tend to colocate with each other and where the colocations exist in the county. More than thirty types of crime (excluding sex crimes) were first reclassified into ten categories, and aggregated into each of the 575 census block groups. Principal component analysis was used at the census block group level to extract components from reclassified crime data to identify the categories of crime that are heavily loaded together and therefore share similar location patterns. Four components were extracted and named contraband, violent, property, and theft. Component scores for each component were mapped and analyzed for census block groups, which helps reveal spatial colocation patterns of crimes in Jefferson County. Temporal changing in crime colocation patterns from 2009 to 2011 was also examined in this study.

Keywords:

• colocation
• crimes
• principal component analysis
• spatial relationship

Additional information

MAX POPE is a graduate of the Department of Geography and Geosciences, University of Louisville, Louisville, KY 40292. E-mail: [email protected]. His research interests include urban crime issues, transportation geography, and GIS applications.

WEI SONG is Associate Professor in the Department of Geography and Geosciences, University of Louisville, Louisville, KY 40292. E-mail: [email protected]. His research interests include location analysis, transportation geography, urban crime issues, and quantitative methods.