Predictive analysis grows as crime-prevention tool
- By David Hubler
- Jan 15, 2013
Although crime rates in many U.S. urban centers have been falling during the past few decades, law enforcement agencies are rightly never satisfied no matter how low the rates go. To push them down further, police departments are adding a new tool, predictive software, to assist their efforts.
Baltimore and Philadelphia, which have used predictive analysis tools for law enforcement, have now added new crime-prediction software that promises “to reduce the homicide rate by predicting which prison parolees are likely to commit murder and therefore receive more stringent supervision,” according to Wired.
The idea is to rely on the software’s predictive analyses capabilities rather than on a parole officer’s instincts and regulations to determine which parolees are most likely to commit homicide.
In a lengthy interview with ABC News in 2010, the creator of the algorithm that powers predictive analytics, Richard Berk, said, “When a person goes on probation or parole they are supervised by an officer. The question that officer has to answer is, ‘What level of supervision do you provide?’ ”
Berk explained that the software simply replaces that kind of ad hoc decision-making that officers already do.
In a transcribed 2012 podcast with IEEE Spectrum, Berk said the software draws on known predictors -- such as a person’s age when he committed a crime, the nature of the crime, the person’s current age and where he lives -- and uses algorithms to build a new set of predictors. A basic example, he said, was that someone who committed an armed robbery at 12 or 13 was likely to be in trouble again at 18. Someone who commits the same crime at 30 isn’t nearly as likely to do so again.
Applying predictive software to criminal behavior has been around for several years, led by the New York City Police Department’s Comparative Statistics process, COMPSTAT, in 1995. Variations are now in use in other major cities including Washington, D.C., San Francisco and Los Angeles.
As GCN reported in 2010, police in Memphis, Tenn., added IBM’s Blue CRUSH (Criminal Reduction Utilizing Statistical History) software to create predictive models by analyzing crime and arrest data.
The CRUSH software reduced serious crime by more than 30 percent, including a 15 percent reduction in violent crimes, over a four-year period, according to IBM. The software enabled the Memphis Police to evaluate incident patterns throughout the city and forecast criminal "hot spots" so they could allocate resources, deploy personnel and increase public safety.
The Philadelphia Police Department introduced electronic crime mapping and forecasting six years ago with its Crime Spike Detector, developed by Robert Cheetham, a University of Pennsylvania graduate and CEO of the GIS firm Azavea, according to Technically Philly.
David Hubler is the former print managing editor for GCN and senior editor for Washington Technology. He is freelance writer living in Annandale, Va.