Budget cuts have taken a toll on the police. Though the SCPD gets 30 percent more calls than it did in 2000, the department has to answer these calls with 20 percent fewer officers. This takes an inevitable toll on the force, but Santa Cruz has come up with a model that may relieve the strain a bit.
It’s called “predictive policing,” and it’s based on something Santa Cruz knows all about: earthquakes. If scientists can use various models to attempt to predict when aftershocks are likeliest, could they use similar models to predict when crimes are likeliest?
Using a sophisticated computer model, the SCPD is able to pinpoint where and when crimes are most likely to occur, based on other events that took place there and identifying patterns based on years of records. As soon as data arrives, the crime bank is recalibrated. Patrol patterns are based on the information to give targeted areas more coverage by on-duty officers.
In one example, two women were caught vandalizing cars in a parking garage. Police were on site because the computer suggested that the particular one-block radius was a high-risk location for car burglaries.
The program itself was a multidisciplinary effort created by a team of researchers, including two mathematicians, an anthropologist and a criminologist. While it has yet to be proven effective empirically, Zach Friend of the SCPD said that burglaries were down 27 percent in July, while the program was operating.
Other police departments are taking notice. Chicago already has a smaller predictive policing program and the LAPD is deciding whether it wants to go that route too. Much of the decision will be based on the success of Santa Cruz’s initial foray into the field. Read more at Santa Cruz Sentinel.
