Catching crime before it occurs
Santa Cruz Deputy Chief Steve Clark has been with the Police Department for 25 years. But there are some things that even experience doesn’t teach. Up until now, he’s been trained to respond to incidents.
“Back in the day, you would ask any police officer what does it take to make your city safe, and the pat answer was, ‘More cops,’” Clark says. “And really, I've described that as kind of primal policing because it’s a primal response. There’s not a lot of thought to it. There’s not a lot of analysis. You’re just going out there and hoping that you get lucky just by sheer numbers.”
These days, Clark is trying to predict what will happen on any given workday. A year ago, he hardly knew what “predictive policing zones” were, but today, they’re an integral part of his job. Clark, like every officer, carries a list of the day’s top 10 “hot spots.”
“What we do is we list those in terms of the highest probability on down,” Clark says. “It’s kind of like our version of Letterman’s Top Ten list, if you will.”
Next to each location is a percentage, based on recent statistics, showing whether a crime is more likely to happen in a car or in a house. That helps officers know where to focus their attention: scanning driveways; watching for open windows on houses; or on people hovering near cars.
At the top of today’s list, is a downtown triple-decker. “Not an ice cream cone,” Clark clarifies, but a parking garage – one that has actually “visited the list quite a bit.” Some residential areas near the boardwalk are also high on the list.
Predictive policing came from an unlikely place. George Mohler, a mathematician at Santa Clara University, and a team of UCLA professors created an algorithm to predict the location of earthquake aftershocks. They used patterns to see how an earthquake originating along one fault will cause additional earthquakes in connected faults. As it turns out, crime waves follow similar patterns to earthquake aftershocks. So they shared the concept with the Los Angeles Police Department.
“There’s nothing quite like getting a team of mathematicians in with a team of police officers and trying to see what you can do,” says Zach Friend, a crime analyst for the Santa Cruz Police Department. He says, usually, cops and stats don’t mix.
They found remarkable results. “Looking at Los Angeles data, they found that certain neighborhoods had characteristics that led to either lower or higher than average crime rates,” Friend explains. “When they boiled that down even further, they found that those crimes clustered not just by neighborhood, but even by time, so that if you had a crime on a certain day, you were much more likely to have other crimes occur within a couple of days or within a couple of hours of when that crime occurred.”
Time, place and type of crime are all that’s factored into the equation. Unlike “hot spot” policing, which is popular in many departments, predictive policing is calibrated on a daily basis. Santa Cruz first implemented predictive policing in the summer of 2011.
“It took quite some time for us to get it set up, and there’s a few reasons why,” Friend says. The department wanted to get officers familiar with the type of data they would be using to inform their patrols. They also felt an obligation to make sure officers knew that the new patrol model was meant to supplement, not replace their talents and intuition. Rather, says Deputy Chief Clark, the model pulls them into areas they might not otherwise spend time in.
“I know that if I go out on a main thoroughfare, I can probably arrest somebody and find somebody who’s doing something,” Clark says. That approach makes sense considering statistics show that's where many crimes and arrests happen. “But they’re not always the ones that are impacting the quality of life in the neighborhoods where people are living or where their businesses are,” Clark says.
Since predictive policing has been implemented, Santa Cruz cops have adjusted their patrols. The model has shown quick and promising results for something police departments usually don’t do: prevent crime. Remarkably, they’ve done so in a time of dwindling city budgets.
“Last year we set a record for the number of calls for service we’ve had in the 160 years of our agency,” Clark says. “We did that with 20 percent less sworn staff, and nearly 30 percent less non-sworn staff, such as the crime analyst, to help out with that system.”
Also last year, the department released an iPhone app for residents to view real-time crime maps, provide crime tips and view photos of wanted suspects. New technology, Friend says, is changing the landscape that law enforcement agencies are operating in.
“Quite frankly, everybody’s cutting police officer positions right now,” Friend says. “So what do you do with less? And I think that using technology like this, leveraging technology is essential to do exactly that.”
Santa Cruz was the first department in the world to implement a predictive policing program. Since last fall, the department has been contacted by about 150 police agencies in the country and law enforcement agencies and news outlets from Brazil, Mexico, Japan, France, Germany and Denmark.