Can We Predict Crime: The case of Burglary in Indio, CA

• In Event: New Findings from the Smart Policing Initiative

Abstract

The notion that policing could become proactive, rather than reactive, is attractive from a crime prevention and a public service point of view. It is widely believed, both inside and outside of law enforcement and criminology that crime itself has enough random elements in its determination that such predictions would be impossible at best. However, borrowing from the new technology industry, we apply the concept of "big data" to the problem of crime prediction. The notion of big data is essentially that having access to large scale data bases and analyzing them for underlying patterns can enhance the understanding and predictability of many behaviors in commerce and Internet activity--so why not crime? The City of Indio, CA and its burglary patterns are analyzed over a 10 year period, a predictive spatial and time series model is developed, and validated against recent observations. An intervention program is concentrated in the predicted hot spots, and is being evaluated to see if those predicted hotspots can be mitigated by Smart Policing based on a predictive, proactive model of policing.

Author

• Robert Nash Parker, University of California, Riverside