Formation processes of the criminological record

• In Event: Predicting crime in space and time: Examining and understanding input, throughput and output

Abstract

Crime forecasting to support real-time policing interventions is based on the idea that timely data allows police to anticipate risk and deploy resources in ways that deter criminal offending. As a matter of convenience, it is often assumed that the data that underlying forecasts are static. It is clear, however, that the empirical features of reported crime can change substantially following initial reporting raising questions about the accuracy of forecasts at the leading edge. This paper borrows ideas from the study of geomorphology to understand the trajectory of the empirical features of crime data from unstable to stable over time. Using crime event data from Chicago collected on a daily basis for an entire year, this work shows that the formation processes of the criminological record are highly regular or law like. It is therefore possible to not only project the survival time for crime records to become stable but also to suggest the types of data corrections that are necessary to make the most robust real-time crime forecasts.

Author

• P. Jeffrey Brantingham, University of California Los Angeles