A Regional Approach to Assessing and Visualizing Spatiotemporal Clustering of Crime Events. (3rd April 2019)

• Record Type: Journal Article
• Authors: Gong, Junfang; Li, Shengwen; Wan, Bo
• Is Part Of: Papers in applied geography. Volume 5:Number 1/2(2019)
• Page Start: 26
• Page End: 44
• Publication Date: 2019-04-03
• Subjects: Localized indicators of spatiotemporal association -- Moran's I -- periodicity -- spatiotemporal autocorrelation -- visualization; Applied human geography -- Periodicals; Geography -- Periodicals; 910
• DOI: 10.1080/23754931.2019.1611625
• Languages: English
• ISSNs: 2375-4931

Abstract

Abstract: Examining the levels of both the spatial and the temporal clustering of crime events raises the analysis of events as spatial patterns at different time periods to understanding them as spatial processes. There have been, however, only limited efforts in doing so, possibly due to a lack of effective methods for estimating, identifying the difference in the levels of spatial clustering patterns, and visualizing the spatiotemporal trends among the events. We first describe a method for examining the periodicity of space–time crime events. Two indexes are extended from the Moran's index for spatial autocorrelation and local indicator of spatial association (LISA) to ones for including spatiotemporal event-region data. We also suggest a three-dimensional approach to visualizing spatiotemporal associations among event regions that enables a better understanding of the spatial processes behind the evolution of event regions. An implementation of the suggested methods that use data of three types of crime events in Akron, Ohio, is reported. The suggested approach is capable of summarizing spatiotemporal autocorrelation as well as identifying spatiotemporal clusters of crime events. Compared to layered LISA, this approach works well, and it allows three-dimensional visualizations of spatiotemporal clusters of events that can help to develop effective policing strategies for combating crimes.