It’s known as the “Moneyball” approach to policing because it harnesses data to predict crime.
For a few weeks now, Seattle police have started their shift by consulting a map that forecasts hot spots for property crime.
The maps use red boxes to highlight areas where officers are more likely to catch a burglar rifling through a home or a car. The boxes are generated by software that analyzes five years of crime patterns to spit out reports highlighting where and when the next break-in could take place.
Officers are then asked to spend two hours of their downtime—the time they’re not out on calls—patrolling the red-box areas.
Seattle Police Sgt. Christi Robbin says deploying officers to hot spots isn’t new; in the past, police relied on crime analysts to make predictions, but the software is two times more accurate.
“The officers really know what the top couple boxes are. They’re common to them; they go to those areas a lot,” Robbin said. “There’s areas where the officers get a box and are like, ‘Hmm, I don’t know this area. This box isn’t familiar to me.’ And it’s a great learning experience in that aspect.”
Robbin says Los Angeles and Santa Cruz have both seen a significant decrease in property crimes since they started using predictive policing a few years ago.
And while it will take months to learn more, Seattle police are already talking about using the software to predict violent crimes like robberies and assaults as early as this summer, the season when violent crime tends to spike in Seattle.
