BI Helps Police Predict, Prevent Crime

Case Study: Data mining vendors SPSS and Information Builders help a Virginia police force take preventive action.

What if IT could help law enforcement outmaneuver the criminal before the criminal got to the crime scene? If this scenario sounds like something right out of the science fiction film "Minority Report," its not.

Using data mining, predictive analysis and business intelligence tools, state and local police in Richmond, Va., are ushering in a new era in crime analysis that hopes to accomplish exactly that.

Combining technology expertise from vendors SPSS and Information Builders, police at the RPD (Richmond Police Department) are applying information-based policy to—they hope—prevent future crime from occurring.

What this means to the general public is, at a minimum, the RPD can respond more rapidly to crime by knowing where to deploy police power and, ultimately, can use information to deter crime by having an increased police presence in the right places at the right times.

"This is a huge step forward for law enforcement," said Rodney Monroe, the RPDs chief of police. "We can capture data, analyze it and put it to use to better serve the community."

For technology partners SPSS, in Chicago, and Information Builders, in New York, the law enforcement application presents a rebranding opportunity that the vendors plan to market to other police jurisdictions.

The application combines predictive analytics from SPSS with Information Builders enterprise BI capabilities, along with an innovative analytic framework developed by RTI International, in Research Triangle Park, N.C.

For 10 years, Colleen McCue, senior research scientist at RTI International, has worked both for and with the RPD, in what she calls "a unique opportunity to apply research in an operational law enforcement setting."

While serving as the supervisor of the crime analysis unit at the RPD from 2000 to 2004, McCue said she used standard records management systems to analyze crime. Then she began using more advanced statistics and modeling techniques to get a better handle on violent crime.

/zimages/1/28571.gifThe city of New York, IBM and Dimension Data collaborate with the NYPD to create a high-tech "crime center" fueled by BI tools and data warehousing. Click here to read more.

"We started to do some interesting things, like linking crimes and motives using advanced statistics and modeling for investigative support," McCue said.

But in the aftermath of Sept. 11, 2001, McCue said she knew the time had come to take her research beyond academic output to making it operationally actionable. Or, in other words, to make police analysis work more like science than fiction.

"By better characterizing crime trends and patterns to predict where crime was likely to occur, the police department could anticipate criminal activity and proactively place their resources," McCue said.

The good news, according to McCue, was that, like most human behavior, criminal activity follows predictable patterns: for example, time of day, week, month and so on; geographic location; event; and weather conditions. In addition, she said, police investigators are natural data miners, sorting through tons of data to make decisions on how to solve crimes.

"Ninety percent of what police departments do is gather, analyze and disseminate data," Monroe said.

Next Page: SPSS enterprise data mining workbench, Clementine.