Share Email Print RSS

Policing via Predictions

Is the Richmond Police Department working with Miss Cleo? Not yet perhaps, but they are trying to predict crimes with the aid of computer software.

Currently, Richmond police turn to predictive analysis technology to foresee possible times and locations of crimes.

Predictive analysis uses a variety of information to determine future outcomes. Utilizing data on former crimes, including location, weather conditions, surrounding events and time, the police use that technology to preempt crime, according to Stephen Hollifield , the department’s information services manager.

"We’re looking at a couple different things when we analyze. We’re looking at all the crime incident information we have," Hollifield said. "That information is fed through a model and is extrapolated to provide an eight-hour window for crime in specific geographical locations."

Police Chief Rodney Monroe signed on with SPSS , the company behind the predictive analysis software, and Information Builders Incorporated , whose software presents the analysis findings through a browser, less than three years ago, and the partnership launched its maiden voyage on New Year's Eve 2005.

On that ever-eventful night, Richmond police tackled random gunfire incidents throughout the city in a preemptive and strategic manner. Using predictive analysis technology, they saw a 49 percent decrease in random gunfire incidents and a 246 percent increase in weapons seized, according to Hollifield.

Hollifield also said predictive analysis is essential for efficient force deployment. On the New Years Eve mission, the department saved $15,000 in overtime pay through innovative deployment strategies.

The companies have frequently discussed goals to enhance the technology with Chief Monroe, according to Bill Haffey , SPSS’s technology director for the public sector.

"Right now this program predicts a tight forecasting timeframe," Haffey said. "The chief would like to project what the city of Richmond will look like in 60- to 90-day periods."

Because predictive analysis is based on past events, Haffey said the models, which are supplemented with GIS mapping and aerial photography, will continuously improve as they gain more information to consider.

"As time goes on, the models will be much more accurate," he said. "It’s going to grow from law enforcement group to law enforcement group."

While the Richmond police is the only law enforcement agency regularly using SPSS predictive analysis, Haffey expects the technology to grow in popularity. He said the Department of Defense made inquiries about their predictive analysis technology for improved troop deployment in Iraq. Moreover, earlier this month the department received the Gartner Business Intelligence Excellence Award , recognizing excellent use of technology in professional settings.

"A lot of applications have value, but we’re looking for something very unique," said Bill Hostmann , Gartner’s research vice president. "This [application] was a compelling case. It was interesting. It was novel. It was creative."

While Hollifield agrees that the program is innovative, he also said it cannot supersede human efforts. The analysis aids the police when they are congruent with the department’s own findings, but the latter is still crucial.

"[Crime analysis] is something that still requires a logical brain," Hollifield said.