Methodology for Evaluating Geographic Profiling Software, Final Report

• NCJ Number: 208993
• Author(s): Tom Rich; Michael Shively Ph.D.
• Date Published: December 2004
• Length: 258 pages
• Grant Number(s): ASP T-037
• Agencies: NIJ-Sponsored
• Publication Type: Report (Technical)

Annotation

This report describes a methodology for evaluating geographic profiling software.

Abstract

Geographic profiling is a criminal investigative technique that provides information on the likely "base of operations" of offenders suspected of committing serial crimes. The base of operations could be the offender's home, place of employment, a friend's house, or some other location frequented by the offender. Conclusions and predictions are based on the crime locations, other geographic information about the case and the suspect, and assumptions about the distance offenders will travel to commit crimes. Interest in geographic profiling software has increased with advances in computer mapping software and the increased use of geographic information systems by law enforcement researchers and practitioners. In order to develop a methodology for evaluating geographic profiling software, an expert roundtable was convened. The first section of this report presents a narrative summary of the 2-day meeting of the panel on August 10-11, 2004. This is followed by background information on various geographic profiling software applications. The evaluation methodology is then described. One component of the evaluation is an assessment of the extent to which the software accurately predicts the offender's "base of operations" by conducting a battery of tests on each software application, using actual serial crime data provided by law enforcement agencies; test results should be compared to results from control methods. The second component of the evaluation is user feedback on the software's utility and ease of use by crime analysts and other staff who have used the software applications. The third evaluation component is a feature analysis that involves comparing a list of features presented in this document, ranked by priority, with the features of the software being evaluated. 48 references and appended transcript of panel discussions