Seattle police using software to predict crime locations
Called “predictive policing,” the software uses data going back to 2008 to predict where crimes are likely to take place on a certain date and time.
Seattle Times staff
Seattle Mayor Mike McGinn and Police Chief John Diaz announced Wednesday that the Police Department has begun using new “predictive policing” software in the city’s East and Southwest precincts as part of an effort to cut crime and eliminate biased policing through analysis of crime data and locations.
McGinn said the technology will allow police to act instead of react in responding to crime while fulfilling a commitment in the city “20/20” plan, which calls for 20 reforms over 20 months to improve policing.
Designed by UCLA in partnership with the Los Angeles Police Department (LAPD) and based on models for predicting aftershocks from earthquakes, the software called predictive policing forecasts locations where crime is likely to occur within a geographic area as small as 500 feet by 500 feet.
A complex algorithm, using data that dates to 2008, predicts where crimes are likely to take place on a certain date and time. Officers will be provided forecasts before their shifts, then use their “proactive time” between 911 calls to patrol those areas, officials said.
“Success will be measured in crime that does not occur,” Acting Lt. Bryan Grenon, a leader in the effort, said of the preventive possibilities.
Officials said the program will bring equal enforcement throughout the city based on data, helping to address biased policing.
The program initially will be aimed at predicting property crimes, because they occur in the largest numbers, police said, with a goal to expand into all five precincts by April. Other crimes, including those related to gun violence, eventually will be added.
According to a Los Angeles Times article on the LAPD’s efforts, predictive policing is rooted in the notion that it is possible, through sophisticated computer analysis of information about previous crimes, to predict where and when crimes will occur.
McGinn and Diaz said the software is estimated to be twice as effective as a human data analysis based on the same information.
Diaz said obtaining the program cost $73,000, along with a $43,000 annual subscription fee.
The LAPD saw a 13 percent drop in property crime during the first year of a test project, the officials said.
McGinn said the predictive policing will help police work together with community members, allowing them to consider measures such as lighting and fences.
The Police Chief, the magazine for the International Association of Chiefs of Police, called predictive policing “the next era in policing.” Time magazine called the Santa Cruz, Calif., Police Department’s predictive-policing program one of the 50 best inventions of 2011, and it was highlighted in a New York Times story.
However, Loyola Law School professor Stan Goldman told National Public Radio in 2011 he was worried authorities could use the data to stop and search innocent people who happen to be in highlighted neighborhoods.
“It may very well end up reducing crime to a certain degree,” he said. “The question is at what cost, at what price?”