Police using analytics to track offenders
- — 22 April, 2011 04:31
Two high-profile murders of college students in North Carolina during 2008 have led to the creation of CJLEADS (Criminal Justice Law Enforcement Automated Data Services), an analytics project that aggregates a slew of disparate data stores and is designed to give law enforcement officials a complete picture of a suspect's criminal history and status.
The suspects charged in the deaths of University of North Carolina-Chapel Hill student Eve Carson and Duke graduate student Abhijit Mahato were on probation at the time of the slayings. Officials believe the centralized repository of court, probation and arrest records and other data CJLEADS provides will help law enforcement workers keep closer tabs on offenders, speed investigations and possibly prevent future crimes.
Work on CJLEADS began in October 2008, with the first pilot going live last year in Wake County, said Kay Meyer, project manager in North Carolina's Office of the State Controller. A statewide rollout began in January, with the goal being to complete the project next year.
So far, CJLEADS has played a role in a number of cases, including the bust of 14 members of a ring that staged car accidents, according to a statement from the North Carolina Department of Insurance.
CJLEADS is a Web-based application based on SAS Institute's Enterprise BI Server, Enterprise Data Integration Server and Enterprise Miner technologies, and is hosted by SAS.
It is funded by North Carolina's General Assembly. Lawmakers initially set aside US$5 million to start the pilot program, and last year slotted a recurring line item for it in the state budget, Meyer said. While the project received nearly $9 million last year, a reduction is expected due to state budget constraints. Still, the project should remain largely on track, she said.
The state is expecting significant financial return on its investment thanks to the amount of time workers will save that otherwise would be spent digging through various databases, one by one.
Still, "most of our savings are really intangible," Meyer said. "It's cost avoidance, versus money we're going to be able to put in our coffers."
So far, user feedback has been solid, according to Meyer. Magistrates "are telling us they're saving 10 to 15 minutes on each first appearance [in court for a defendant]," she said.
While CJLEADS serves a specialized purpose, the state has learned some lessons that may apply to just about any major BI (business intelligence) or analytics project.
For one, end-user input into the system's design was key. The state worked with panel groups composed of police, court and other officials to gauge their needs, with some valuable results.
Police asked for a number of "hot keys" that would allow them to rapidly perform tasks such as a license plate inquiry while in their vehicles, Meyer said.
They also pointed out that text rendered as black-on-white would cause the screen to be quite bright at night, illuminating an officer's face. This could cause night blindness as well as let suspects get a closer look at what the officer is doing, she said. As a result, SAS developed a "night vision" version of the software that uses a black background with white text, she said.
The state also had to overcome "huge" and somewhat unforeseen obstacles around data governance, Meyer said. "It took a lot of work to get everybody on board with the sharing of data and the control and access of data. ... Some of them took longer than we expected. Everybody's just trying to adhere to their own policies."
Chris Kanaracus covers enterprise software and general technology breaking news for The IDG News Service. Chris's e-mail address is Chris_Kanaracus@idg.com