Santa Cruz PD innovation 'predicts' crime hot spots
SANTA CRUZ, Calif. It's more than a hunch that there's a rise in car break-ins as people go holiday shopping and don't put their valuables out of sight. Santa Cruz police are using high-level math to predict where crimes will occur before they even happen.
"On average, we're somewhere between 16, 17 percent reduction in vehicle thefts and any time of burglary -- be it against a home, a business or a car," said Zach Friend with the Santa Cruz Police Department.
The predictive policing program was devised by Prof. George Mohler at Santa Clara University by using algorithms. It's like predicting aftershocks following an earthquake: Maps can be produced showing officers where a break-in will occur, using seven years of crime statistics and current data.
"There's about a 2 percent chance that within this exact box today, there will be one of these crimes," said Friend, showing a display with a map. "They look like boxes, but they're people. This is somebody who didn't have a crime occur to them."
At the start of each watch, officers are given a list and a set of maps showing 500-by-500 foot sectors where the probability is highest of a crime.
In a typical hour, officers spend 50 minutes handling calls. The rest of the hour is focused on the predicted crime areas. This focuses what they will do in that 20 minutes, to come through these specific locations whenever they can, and that presence helps deter crime.
TIME Magazine has singled out the five-month-long test as one of the top 50 innovations of the year. The program is web-based and fried.
Friend says there's one drawback though.
"The fact that I have to be in seven days a week to do this individually," Friend said.
Success of the program hasn't gone unnoticed: Los Angeles police implemented the same math-based preventive policing last week.