Each morning at the Real Time Crime Center in Memphis, Tennessee, police officers scan walls of video feeds from hot spots around the city while computers spit out the latest crime predictions. A red dot flashing on a map signals that a crime may happen on that block soon. If a commanding officer thinks the software is correct, he’ll send a patrol ahead of time to catch the criminal red-handed. Better yet, the police presence may prevent the crime from happening at all.
Memphis police director Larry Godwin assures the public that this isn’t a real-life version of Minority Report. In Steven Spielberg’s sci-fi thriller, psychic mutants immersed in goo foresee criminal activity so that Tom Cruise and his “precrime” officers can arrest would-be suspects before they act. In Memphis, no one is arrested preventively. But the software does aim to forecast burglaries, drug sales, gang violence, and other illegal acts before they take place, says Godwin.
The predictive software, which is called Blue CRUSH (for “criminal reduction utilizing statistical history”), works by crunching crime and arrest data, then combining it with weather forecasts, economic indicators, and information on events such as paydays and concerts. The result is a series of crime patterns that indicate when and where trouble may be on the way. “It opens your eyes within the precinct,” says Godwin. “You can literally know where to put officers on a street in a given time.” The city’s crime rate has dropped 30 percent since the department began using the software in 2005.
Memphis is one of a small but growing number of U.S. and U.K. police units that are turning to crime analytics software from IBM, SAS Institute, and other vendors. So far, they are reporting similar results. In Richmond, Virginia, the homicide rate dropped 32 percent in one year after the city installed its software in 2006.
Some of the funding for such setups is now coming from the National Institute of Justice (NIJ), the R&D arm of the U.S. Justice Department. Other funding is coming from nonprofit groups. This year, the nonprofit RAND Corporation teamed up with the Chicago police department to apply predictive analytics to gang behavior.
The increase in funding may help push more big police departments to take up such initiatives, says Jeffrey Brantingham, an associate professor of anthropology at the University of California, Los Angeles, who leads a research team of UCLA academics and Los Angeles police officers that is seeking a $3 million NIJ grant to test predictive policing models.
Brantingham says his approach is less about adapting the software for L.A. and more about spotting predictable patterns of universal human behavior. “People tend to utilize their local environment, so they don’t travel long distances to do things like buying milk,” he says. In the same way, he adds, “most burglars victimize places that are very close to where they live, or where they work, or where they hang out.”
With demand on the rise, IBM is betting big on predictive analytics software. Over the past four years, the company has invested $14 billion in more than 24 acquisitions to expand its analytics division, according to Robert Reczek, an IBM communications executive. Reczek says that more than 200 IBM mathematicians focus exclusively on analytics.
Police departments aren’t IBM’s only projected clients. The software also has applications ranging from preventing Medicare fraud to spotting phony university admissions data to detecting information leaks within federal agencies, says William Haffey, director of sales engineers at SPSS, the division of IBM that developed the software.
Sometimes, these programs pick up unpredictable warnings. For example, police might catch an uptick in shoplifting at the local mall “if it turned out that if it had been raining for three days, and it stopped raining and it happened to be the Saturday of a three-day weekend,” says Haffey.
But as promising as predictive analytics sound, simply buying the software is no panacea. After the Florida Department of Juvenile Justice purchased some from SPSS a few years ago, the job of projecting next year’s crime rates shrank from a day to a few hours, says the department’s chief of research and planning, Mark Greenwald. Still, Greenwald would need a much larger budget to start making the kinds of detailed predictions seen in Richmond or Memphis.
“I think it has been useful, at least for our general trend forecast, and from my perspective it’s very easy to use,” he says. “But it has a lot of functionality that I haven’t been able to tap into yet, because of staffing issues.” In other words, the technology won’t do much good if departments don’t have enough technical staff to keep feeding the model with the latest data—or if they don’t have enough cops to hit the hot spots when warnings flash.