In the movie “Minority Report,” crime fighters were able to visualize crimes before they happened, getting enough details to stop the crimes in progress and arrest the would-be perpetrators.
Reality today isnt quite so dramatic. But law enforcement agencies and corporations are turning to developers such as SPSS Inc. and SAS Institute Inc. and using predictive analytics gleaned from data mining to gain greater insights into solving and in some cases, even preventing crime.
The Richmond (Va.) Police Department has deployed Chicago-based SPSS Clementine data mining software to help snuff out crime before it happens, prevent property crimes from escalating into more violent crimes, and even gain insights into how the drug trade operates.
“Its as close to a crystal ball as were gonna get for a really long time,” said Colleen McCue, program manager for the Crime Analysis Unit at the department. “It truly is a paradigm shift.”
The department scored recent successes using the software, including a crackdown on the firing of guns into the air on the Fourth of July. “We put resources out in locations where historically weve had problems with the illegal use and discharge of weapons, and we were able to get a lot of illegal guns off the street,” McCue said.
The software has done more advanced analysis too. Three times in the past 12 months, the Richmond Police Department used the software to accurately predict that property crimes in progress could escalate to a violent crime.
“We flagged what would normally be a standard property crime and rolled out a more aggressive response and had the right force in place at the right time,” McCue said. “In all three cases, we pulled a violent offender or sex offender off the streets.”
The software also helps to pinpoint crime by type and area to help with tactical deployments of personnel, especially as the department, like so many others, has to cope with budget crunches.
The Police Department has used Clementine for the past year and a half, after using SPSS base analytics platform before that. Clementine has enabled the department to deploy the software for use around the clock by nontechnical users, said McCue, herself a statistician.
“Were starting to get towards 24-by-7 crime analysis. Most of us analysts work 9 to 5. Most crime happens on evenings and weekends,” she said.
Homeland Security Implications
The technology also has homeland security implications because its able to analyze telephone and Internet usage records, though McCue was reluctant to talk about specific ways the department is using it in those areas.
Homeland security is also on the mind of Wachovia Corp. The banking and financial services company has to comply with terms of the USA Patriot Act that require it to have a better handle on possible money laundering and terrorist financial activities. By next year, Wachovia, of Charlotte, N.C., plans to deploy SAS Anti Money Laundering solution to find patterns in transactions that could indicate suspect activity.
“The Patriot Act is very broadly written,” said Bill Langley, executive vice president and chief compliance officer at Wachovia. “What it says to me is that if we purchase a system to assess compliance, we have to purchase one thats flexible and advanced and go with a vendor thats going to be in business for a while.”
Wachovia will use the technology from SAS, of Cary, N.C., to look for patterns of activity that would indicate money laundering as well as unusual customer behavior that wouldnt fit a customer profile, such as a small business conducting unusually large transactions.
Wachovia also needs more powerful analytics that could tie multiple transactions from multiple customers together to discern patterns.
“If you look at the 9/11 terrorists, a lot of the transactions they had processed seemed very innocuous on their own,” Langley said. “But if you can link behaviors together and see a group of people making transactions at ATMs in three or four different states over a few weeks, maybe its something, [or] maybe theyre just on vacation.