Uncategorized

Predictive Policing in Practice

news2GT rides along with the SCPD to test acclaimed new policing program

Is it realistic to predict crime and stop it before it happens, or just a science fiction-esque impersonation of Minority Report? This was one of the questions I pondered on a recent ride-along with Santa Cruz Police Department (SCPD) Deputy Chief of Police Steve Clark. Little did I know I was about to witness a poignant first-hand example of how the “predictive policing” method can be successful.

The department’s adoption of the technology-based predictive policing program has received national and international attention in recent months, with a reporter from Popular Science jetting to Santa Cruz to see it in action and a nod from TIME magazine, which named the fledgling program one of the Top 50 Innovations of 2011. GT set out with these accolades in mind to see for ourselves the result of the department’s six-month evaluation of the predictive policing program, which wrapped up in January.

As we set off in the black and white patrol car, Clark launches into an explanation of how they came to use the technology, and how it has worked out.

The SCPD was the first-ever police department to implement the program in July 2011. They did so just as the city was on track to set the record for most reported burglaries in local history. Between 400 and 500 burglaries occur in an average year in Santa Cruz, but according to the SCPD, 2011 could have seen far more. (They ultimately responded to 568 burglaries and 185 stolen cars in 2011.)

Over the course of a six-month trial evaluation of the predictive policing program between July 1, 2011 and January of this year, burglaries decreased by 11 percent.

According to SCPD spokesperson Zach Friend, the department considers the six-month experiment a success.  

“After implementation of the program, we ended up higher than average but below the historic number [of burglaries] we were approaching,” he says.

Predictive policing employs large sets of data and a complex algorithm to forecast when and where future crimes are most likely to occur, and how officers could be deployed preemptively to stop them.

The technology was the brainchild of Santa Clara University mathematician George Mohler, who introduced it to the SCPD in October 2010. Mohler based the technology’s algorithm on one used by seismologists to predict earthquakes and their aftershocks. He found that such crimes tend to cluster and spread in a way that is similar to tremors after a large quake.

The algorithm targets property crime, including home burglaries, car break-ins and vehicle thefts. Santa Cruz does not have enough violent crime data  for the system to predict those.

As Clark and I cruise past the Santa Cruz Beach Boardwalk, he explains that predictive policing is designed to make use of patrol officers’ unobligated minutes when they are not answering calls. It’s no replacement for intuition and experience, he adds, but rather an enhancement.

“You still have to take those observation capabilities out there in the street with you,” he says.

He hands me a stack of papers that are also given to SCPD officers prior to their patrol shifts. The papers detail the predictive policing program and contain statistics regarding the most likely crime areas in town on a given day and maps of the crime-prone areas, or “hot spots.”  

Clark pulls over near Seabright Avenue and asks me to pick one of the red exclamation points on the largest map. Each point corresponds with a 500-by-500 foot “hot spot.” I choose one pointing to a neighborhood in midtown, across the street from Whole Foods on Soquel Avenue.

When we arrive at the hot spot, he tells me, “One of the first things I like to do when we go into one of these areas is to drive around a little bit and take the temperature of the area—who are the eyes and ears of the neighborhood today?”

The street appears quiet. Clark explains that we are looking for any signs of suspicious activity, such as window screens that have been removed, front doors left askew, unlocked cars, or empty cars with their windows down.

 At a stop sign on the corner of Cayuga Street and Soquel Avenue, Clark’s eyes narrow toward a parked car with a patchwork paint job and three silent passengers. We circle the block once more to come up and park behind this “suspicious-looking vehicle.”

“I’m going to check these guys out,” he says.

 Minutes later, I sit in the patrol car as Clark searches two of the car’s passengers. His voice crackles through the car’s radio as he dialogues with headquarters, and I listen as the dispatcher confirms that one passenger is on parole in Mendocino County with a warrant out for his arrest, and that the other is on probation for possession of stolen property and has two warrants out for her arrest in Mendocino and Santa Cruz counties.

Since our car is not equipped to detain arrestees, a back-up police car arrives. The two arrestees are handcuffed and taken away.

Clark opens the patrol car door and leans toward me, displaying the evidence he collected when searching the vehicle: one methamphetamine pipe and 14 tiny plastic bags with equally measured white grains of what he suspects is methamphetamines, or “crystal meth“—evidence enough to make a convincing drug sale case against the arrestees, he adds.

A chemical color-test back at the police station verifies that the confiscated drugs are methamphetamines. Clark weighs each of the little drug-filled bags, and discovers that they each weigh exactly the same amount, which he says confirms the likelihood that these bags were for sale.

Recounting the day’s events, Clark tells me he also confiscated a flashlight, police scanner, butterfly knife and mask when he searched the two suspects.

“That was predictive policing at its finest,” he says. “That’s everything that you want to look for when you’re doing a predictive policing check because those are the people [who] are gonna be your burglars.”

As SCPD continues to utilize predictive policing, they hold daily meetings to assess ways in which to fine-tune the program.

Photo: Jesse Clark

To Top