This is a summary of a study that investigated alternative methods of forecasting major crimes one month ahead for fixed area units (precincts and square grid cells) that comprise a jurisdiction.
This research is among the earliest attempts to determine the feasibility of crime forecasting, including both extrapolative methods and leading indicator models. The research compared common police practices with simple, widely used forecast models through a state-of-the-art experimental design and extensive police data from Pittsburgh, PA. Although initial results are promising, the researchers advise that they do not know whether crime forecasting will ultimately be accurate enough for use by police. Still, in order to provide a perspective on how crime forecasting could fit into policing and crime mapping, the study developed a use case scenario (a fictional story). It is the target that the researchers envision for research on crime forecasting. After providing the scenario, this report reviews the alternative approaches to short-term forecasting in more depth. It then describes the Pittsburgh case study for evaluating forecasts. This is followed by a description of the experimental design for assessing forecast accuracy, along with the results. Based on research findings, six recommendations are offered for police and four recommendations are provided for researchers. 11 exhibits and a glossary
Downloads
Related Datasets
Similar Publications
- The Experience of Social Distancing During the COVID-19 Pandemic: A Qualitative Lens on Variability in Compliance
- A Practical Guide to Interviewing Potential Human Trafficking Victims
- Optimizing Public Surveillance Systems for Crime Control and Prevention: A Guide for Law Enforcement and their Municipal Partners