Hot Spots of Crime: Methods and Predictive Analytics

Abstract

As part of a contemporary look at the intersection of geography and crime, the current chapter offers a new conceptual definition of predictive policing, an overview of some of the most commonly used methods to identify crime “hot spots” retrospectively, prospective hot spot analysis, and predictive policing analytics. Police and researchers have mapped crime and disorder for decades, consistently finding that these incidents clusters in areas commonly referred to as hot spots. Understanding where and when crime and disorder cluster, both spatially and temporally, provides vital information necessary for community leaders to design and implement effective crime-reduction strategies and community-safety initiatives. Until recently, however, researchers and analysts have operated retrospectively—using incident location information to describe historical crime patterns through the use of data visualization techniques or to design reactionary policing strategies based on the assumption that historical crime patterns are reliable indicators of future problem areas. Unlike traditional crime hot spot analysis, predictive analytics identify patterns in crime data that relate to criminal activity in the near future. The aim of this approach is to exploit patterns identified in data dynamically and to prevent crime through proactive resource allocation. This approach is distinct from how crime analysis and intelligence-led policing has historically been conducted.