Understanding Spatiotemporal Patterns of Multiple Crime Types with a Geovisual Analytics Approach

Comprehensive crime data sets have been collected over time, which contain the location and time of different crime types such as aggravated assault or burglary. To understand the patterns and trends in such data, existing mapping and analysis methods often focus on one selected perspective (e.g., temporal trend or spatial distribution). It is a more challenging task to discover and understand complex crime patterns that involve multiple perspectives such as spatio-temporal trends of different crime types. In this Chapter we used a data mining and visual analytics approach to analyze the crime data of Philadelphia, PA, which has all the crimes reported from January 2007 to June 2011. Specifically, the adopted approach is a space-time and multivariate visualization system (VIS-STAMP) and the analysis examines the spatial and temporal patterns across six crime types, including aggravated assault, robbery, burglary, stolen-vehicles, rape and homicide. The geovisual analytic tool provides the capability to visualize multiple dimensions simultaneously and be able to discover interesting information through a variety of combined perspectives.