This Small City’s Police Department Builds An App, Nabs Big Data To Find And Fight Bad Guys


Neal Ungerleider

March 26, 2014

Rochester, Minnesota (pop. 109,000) is a small city perhaps best known as the home of the Mayo Clinic.

But that doesn’t mean the town’s police department is not busy–and connected. Last year the Rochester Police Department (RPD) decided to build an app to keep its officers’ 160 work smartphones connected and in communication when on the job. Like other tech-savvy departments, including the New York Police Department and the San Francisco Police Department, chief of services for the RPD Captain Tim Heroff decided his cops needed identity-researching apps on their phones.

The RPD mobile app keeps cops in communication, and also uses license plate numbers as a portal into a vast trove of intelligence on vehicles and their histories. This gives officers in the field more information about a car and the people associated with it than the driver themselves may know.

[The app] does a very quick analytical deep dive look into anyone who has been associated in any way, shape, or form with this vehicle. All of that information is then delivered within five to 15 seconds.

In January officers began beta testing the new app, firing it up, for example, during traffic stops. Officers punch a license plate number into the app, which runs queries in a system called InfoSphere Identity Insight which the department leverages. Heroff, who is closing in on 30 years at the RPD, says the app then “does a very quick analytical deep dive look into anyone who has been associated in any way, shape, or form with this vehicle; (that information) goes through a filter to find pre-identified prolific and serious offenders, anyone with a warrant, and anyone on probation. All of that information is then delivered back to officer on the street within five to 15 seconds.”

Heroff told me that this information is used to see “non-obvious relationships and associations” between suspects. “If a parole officer or investigator knows who they are dealing with and the fact that it is one of our identified priority offenders, they have the opportunity to have a heightened sense of situational awareness and know better how to provide for increased officer safety,” he says.

Essentially, the RPD’s app lets police officers conduct background searches on vehicles and their drivers in near-real time, and to quickly generate social network graphs of anyone associated with a particular car, motorcycle, or truck. InfoSphere Identity Insight is one of a series of applications made by IBM and other large vendors such as Microsoft for law enforcement to conduct searches on the fly. The massive computing power of smartphones and easy access to server power via cloud computing have changed law enforcement.

Rochester’s app is built on top of the IBM platform. The RPD follows a law enforcement philosophy known as “intelligence-led policing,” which is best known in its manifestation in New York City, where the data-driven CompStat system focuses on minor offenses like turnstile hopping or public drinking. The big idea is that individuals who commit minor offenses are more likely to commit major offenses down the road. Heroff says that in the RPD’s interpretation of intelligence-led policing, a small sub-segment of all offenders are responsible for the majority of crime. InfoSphere Identity Insight and the RPD app, then, help police zero in on people they regard as likely perpetrators.

The town’s residents have what Heroff euphemistically calls ‘higher than average expectations in how we use technology.’

Rochester is a generally quiet city with an educated populace that centers around a prominent medical institution and a major IBM production facility (hence the IBM connection). The city is a good fit for the tech-minded Heroff. Compared to many other police departments, he says, Rochester’s data is stored in a relatively small number of databases. Heroff added that the town’s residents have what he euphemistically calls “higher than average expectations in how we use technology”–as well as a supportive city council.

The app isn’t the only use of cutting-edge intelligence tools the RPD uses. Heroff said the department generally prefers analytics systems that track specific criminals as opposed to the geography-centered GIS tools favored by many other police departments. Heroff said that 14-15% of Rochester’s criminals are responsible for 50% of the city’s crimes; the RPD then uses InfoSphere Identity Insight to cross-compare data they already have in records. This means that his colleagues can slice and dice info from Rochester PD databases, the county sheriff’s office, outstanding warrants databases, parole databases, and (legal in Minnesota according to state law) public utility records.

While such data-crunching has many benefits, it also creates an environment with very real risks of misuse or unintended consequences. Merging information from different databases carries the risk of mixing up records or accidentally importing corrupted or incorrect data. Algorithm-driven analysis platforms also run the risk of considering innocent people as possible criminals. On the plus side, it can cut costs for police departments, saves taxpayers money, decreases the time of pull-overs, and perhaps even increase public safety.

It’s too soon to gauge the app’s effectiveness in bringing down crime rates, and IBM doesn’t offer any metrics on InfoSphere Identity Insight’s success, saying only that “it mines the overall information landscape in real time, bringing to light all of the associations and occurrences involving the same person or a group.” A look at the city’s crime rates between 2000 and 2012 shows that most major crimes have not fluctuated much over that time. The RPD first began using IBM’s system in 2012.

The immediate future of the RPD will only be more tech-propelled: The application will be rolled out department-wide in the first week of April, and the Rochester Police Department will be issuing Samsung Galaxy Note 3 phones to all officers, loaded with the proprietary software.

Great ! Thanks for your subscription !

You will soon receive the first Content Loop Newsletter