Powerful new tool in Minneapolis creates virtual city

Information about every aspect of life in Minneapolis will soon be combined into a virtual city at the fingertips of city employees, giving them a powerful new tool to do everything from predicting crime to preventing traffic jams.

Departments across City Hall now track everything from licenses, city events and landlords to the locations of bars and police cars. A partnership with IBM will soon allow them to join and visualize that data in new ways to learn more about what's happening across the city, past and present.

It could eventually alert police to "hot spots" before they become obvious, for example, or predict crime using 311 "suspicious activity" calls. Other potential uses, city officials say, include finding youth at risk to violence, tracking food stamp use, identifying elderly susceptible during heat waves and spotting vacant properties earlier.

"The intention of this program is to help coordinate information between and among the various departments to help make our public services more efficient and more useful to the residents of Minneapolis," Council Member Betsy Hodges said Tuesday.

Privacy experts warn that such a powerful tool could infringe on people's rights, however, if patterns mistakenly lead the city to target innocent citizens or employees misuse their access — currently a problem statewide with Minnesota's driver's license database.

City officials say they're taking pains to ensure that names and other identifying information won't be included. "The intention here is not to play Big Brother," said Hodges, who chairs the council's Ways and Means Committee, which approved some of the conditions Tuesday.

The ambitious project is made possible by a three-year, $2.8 million agreement with IBM that city officials are preparing to sign next month.

"We're interested to see exactly how [employees] use this," said Otto Doll, the city's chief information officer. "And we've got a feeling they're going to find value in ways that we don't even imagine today."

Cities across the globe have coupled with IBM to solve problems under the company's "Smarter Cities" banner, but Minneapolis' project will be unique. The core goals will be to spot anomalies, determine causes and effects, find hot spots and discover patterns among many data sets.

What makes most of it possible is a tool called the Intelligent Operations Platform, an interface that allows data to be shared, mapped and mined across city departments.

Lt. Jeff Rugel, who is testing the tool for the police department, said it will give precinct lieutenants an easy method to visualize — geographically and temporally — crime that has occurred on the previous shift, beyond just reviewing police reports. That task now often requires the help of crime analysts. They can go back weeks or months, isolate certain times of day, map things relative to bars in an area, or observe the nearby permits. That means officers can "very quickly follow up on a hunch," Rugel said.

In the future, it could be configured to catch certain crime patterns, such as two burglaries happening near each other within a certain time frame. "You don't want to spot the pattern when it's obvious," Rugel said. "What makes it obvious is more victims."

Doll imagines it could also be used to provide staff with daily updates documenting the status of the city's 13 wards, pulling information on violent crime, businesses statistics, infrastructure activities and residents' complaints.

He envisioned it could show whether the presence of a police car on a city street reduces criminal activity, using GPS data on squad cars and police reports. Or, using a combination of variables, they could determine which at-risk youths are in need of intervention or mentoring, Doll said in a September speech.

But the consolidation of all that data also worries local privacy advocates, who say the system needs public accountability to protect against profiling and invasions of privacy.

Among the "potential uses" departments identified in the December presentation were "chronic offenders geographic restrictions," "identify households prone to domestic violence," "monitor problem landlords" and "track homelessness movement."

Doll said they made a decision not to allow personally identifiable information into the system, a line he acknowledges is "murky." Named data and Social Security numbers are off limits, he says, but addresses are generally not. They may strip identifiable features out of data sets using a key only some people could decipher. His office will review all requests to add data.

"Information is power," said Rich Neumeister, an open government expert. "And the public needs to know when their public resources are being used to make the government even more powerful ... with the use of analytics, intelligence gathering, intelligence analysis."

Chuck Samuelson, executive director of the Minnesota branch of the American Civil Liberties Union, said being able to drill down information allows governments to see and act on things that were not otherwise apparent. "But when they're wrong, they're spectacularly wrong," Samuelson said. "People go to jail when they shouldn't be in jail. Or people get investigated when they shouldn't be investigated."

Doll said the system will initially rely on existing city data sets, but may eventually include county, state or federal data.

Just how that will play out in the police department remains to be seen. Rugel said, "Someday I want 6,000 more data fields in there. Because you can never have too much to search."

Later, he added, "I think someday that would be a benefit if you could say, 'Well, we've had some burglaries. Who in this area has a history of burglary arrests?' "

Eric Roper • 612-673-1732