Red boxes in West Seattle indicate property crime hot spots, based on historical crime data collected by a Predictive Policing computer program. Southwest Precinct officers will use the information to patrol areas they predict criminals will be working on a given day.

UPDATE: 'Predictive Policing' comes to West Seattle

Through data analysis of likely crime hotspots, SPD hopes to enhance prevention

To supplement a cop’s intuition, the Seattle Police Department is using historical crime data in an attempt to predict where crimes will occur before the criminals even show up.

Seattle Mayor Mike McGinn announced “Predictive Policing” software has been deployed in two precincts, including the Southwest Precinct of West Seattle, as of Feb. 27, with the promise that “This technology will allow us to be proactive rather than reactive in responding to crime.”

The software, developed by UCLA and the LAPD in Los Angeles based on models that predict earthquakes, will attempt to forecast where crimes might occur based on data analysis of logged crimes since 2008, according to the Mayor’s office. Police will use that data to determine where they patrol, searching for burglars and thieves in our neighborhoods.

“The Predictive Policing software is estimated to be twice as effective as a human data analyst working from the same information” SPD Police Chief John Diaz said in a statement. “It’s all part of our effort to build an agile, flexible and innovative police department that provides the best service possible to the public.”

The software is initially only being deployed in the Southwest and East Precincts, and only focusing on property crimes (burglaries, car prowls, auto thefts). McGinn said it will be expanded to all precincts and eventually look at violent crime as well.

The focus on property crime is significant in West Seattle, where home burglaries, car prowls and auto thefts far outstrip violent crime, like robberies, on a weekly basis.

Technically, “Predictive forecasts the locations where crime is likely to occur, down to a geographic area as small as 500 feet by 500 feet. It works by entering all crime and location data dating back to 2008 into a complex algorithm that generates a prediction about where crimes are likely to take place on a certain day and time. Officers are provided with these forecasts before beginning their shifts, and are assigned to use their ‘proactive time’ between 911 calls to patrol those areas.”

As shown in an SPD PowerPoint presentation on the technology, officers will have a map that displays red boxes indicating potential hot spots for crime on a given shift.

Removing bias from the equation
The implementation of data-driven analysis is part of the SPD 20/20 Plan laid out by the City of Seattle to correct issues in the department identified by a Department of Justice review in 2012, including concerns over racial bias.

By only collecting data on the type, location and time of crimes since 2008 and not taking into account personal information on who committed the crimes, the city believes this software “meets community expectations for privacy and addresses concerns about bias.”

“We anticipate that this software will help us in our work to eliminate institutional bias in the Seattle Police Department” SPD Assistant Chief Mike Sanford said in a statement. “With Predictive Policing, we’re sending officers right to where we know the crime is, helping to take any unconscious bias out of the equation.”

Encouraging statistics from California
According to a 2012 CBS News report on the same software being used by the Foothill Division of the Los Angeles Police Department, the results of Predictive Policing have been significant. The foothill division of LA is, much like West Seattle, primarily made up of working class neighborhoods.

“Sixty-five percent of our crimes are burglaries, grand theft auto and burglary from a motor vehicle,” LAPD Captain Sean Malinowski told CBS, explain how the red “hot spot” boxes on their maps represent areas most commonly hit.

The Foothill Division started using the software in November of 2011 and, after five months, reported burglaries were down 33 percent.

Time will tell how effective Predictive Policing will be in West Seattle, but encouraging statistics from elsewhere combined with the robust culture of block watches on the peninsula just might send property criminals scurrying to other parts of the city and county in 2013.

We encourage our readers to comment. No registration is required. We ask that you keep your comments free of profanity and keep them civil. They are moderated and objectionable comments will be removed.