Space–Time Clustering of Crime Events and Neighborhood Characteristics in Houston

Abstract

Spatial–temporal interaction analysis is employed to identify repeat and near-repeat patterns of crime in time and space. Most research to date addresses burglary and shooting incidents. Using the Knox method for space–time interaction, this study analyzes crime data in 12 “super neighborhoods” located in Houston’s crime-heavy southwest quadrant to explore spatial–temporal clustering of three types of crime, namely, residential burglary, street robbery, and aggravated assault. The findings suggest that each type of crime event has a unique clustering signature. Residential burglaries show significant space–time clustering in a relatively longer time range (up to 90 days) and distance interval (up to 1.55 miles). In contrast, street robberies present significant clustering only up to 6 days and a quarter of a mile. For aggravated assault, the clusters of pairs occur within the interval of 7 days and within a little more than 1 mile of an initial assault. Examination of the socioeconomic characteristics of the neighborhoods indicates that crime events cluster more often in low income and racially/ethnically diverse neighborhoods. Significant spatial correlations of crime clusters are detected. The findings offer insight into potential suppression of crime events that are time and space correlated.