Testing Geospatial Predictive Policing in a Medium-Sized Police Agency

The promise of predictive policing is that police agencies can intervene proactively in problem areas and prevent crime and violence from occurring. Predictive policing is based on the premise that crime can be forecasted or predicted accurately for small enough geographic areas and on reasonable enough time frames to enable police agencies to deploy officers to potential problem areas.

Although many tools are available, enabling proactive approaches to policing (e.g., hot spot analysis, network analysis, data mining), there is a clear need to link appropriate theory to data to enable predictive policing strategies. There is also a need to uncover the mathematical and statistical models that underlie those strategies.

Researchers at Justice & Security Strategies, Inc. (JSS), will work with the Columbia, S.C., Police Department and consultants with the University of South Carolina on a project to meet four goals:

  • Link routine activities theory and appropriate measures (data elements) to the practice of predictive policing.
  • Determine the accuracy of various predictive algorithms, including traditional hot spot analyses, regression-based analyses and data mining algorithms.
  • Implement the predictive policing process in the Columbia Police Department.
  • Evaluate the effectiveness of the predictive policing process using a randomized control trial and regression-discontinuity design.

In pursuing these goals, the project asks a number of research questions: What is involved with the entire process of predictive policing? What does it take to implement? What are the underlying theories that guide it? What data are needed? What type of software or hardware is necessary? How does predictive policing “work” in the field? What is its practical utility? How do we know that it is effective? How do we translate what we know into useful, useable strategies and tactics on the street to prevent and reduce crime?

In addition to written reports and other deliverables to NIJ, JSS researchers will write articles that explain the complexities of predictive policing in plain language to promote translating research with significant ramifications for police and policymakers into practice.

See grant details for "Linking Theory to Practice: Testing Geospatial Predictive Policing in a Medium-Sized Police Agency," Justice & Security Strategies, Inc., NIJ grant 2013-IJ-CX-0054.

Date Created: June 9, 2014