Predictive Policing Symposium: The LAPD Experiment

Charlie Beck
Chief, Los Angeles Police Department

Note: The Los Angeles Police Department is one of seven agencies that were awarded a planning grant from the National Institute of Justice. For more information on the other six, see the summary of the Discussion on the Predictive Policing Demonstration Projects and Evaluation.

Police Chief Charlie Beck discussed how the LAPD has used COMPSTAT, intelligence-led policing and problem solving to reduce crime over the last eight years.

Predictive analytics have been used by businesses to determine sales strategies. Wal-Mart, for example, analyzes weather patterns to determine what it stocks in stores. The results indicate that Wal-Mart should overstock duct tape, bottled water and strawberry Pop-Tarts before a major weather event. The Pop-Tarts represent a “nonobvious relationship” and Beck noted there are many of these relationships in law enforcement that can be explored with predictive policing.

Beck emphasized that predictive policing is not meant as a replacement strategy, but one that will build on the successes thus far.

Colleen McCue illustrated the potential of predictive policing. For example, every New Year’s Eve, Richmond, Va., experienced an increase in random gunfire. The Richmond Police Department looked at the when, where and what of these cases and found that the gunfire was actually limited to certain areas of the city within a two-hour timeframe. McCue said this was not a complicated analysis, but by looking at the data with a different eye, the department deployed resources more efficiently. The result was a 47-percent reduction in random gunfire, a 246-percent increase in seized weapons and a savings of $15,000 with 50 fewer officers deployed.

Dr. Sean Malinowski explained that the predictive policing movement began in the LAPD seven years ago on a grassroots level. Today, the LAPD has significant momentum for its predictive policing efforts in counterterrorism, robbery and social network analysis.

At first, crime was summarized annually. The analysis improved, and the department began to look at data monthly, then weekly and eventually in real time. The department expects to forecast crime by 2010, resulting in smarter policing overall.

Capt. Justin Eisenberg summarized three of the LAPD’s predictive policing projects that are in different stages of development:

  • Debriefing Project: Officers debrief arrestees to obtain information unrelated to the crime for which they were arrested. Officers will capture anything that arrestees are willing to provide, including information on social networking, terrorism, bomb making and more. This information will be used to identify nonobvious relationships. Currently, the department is finalizing a Web-based database for the project.
  • Social Networking Analyses: Specific to gang investigations, this project involves adding an identification software tool for social networking analysis to the work that the University of Cincinnati and the Cincinnati Police Department have done. The LAPD hopes to identify gang decision cycles to remove main players from the street.
  • Gang Homicides: This project involves mapping the details of gang homicides to more effectively predict future murders. 
Date Created: December 18, 2009