March 3, 2003

tephen Sacks draws down a menu on his laptop computer. He clicks his mouse, and a map of New Britain on the screen is suddenly broken into 70 census blocks. As he clicks on another link, the map fills with multi-colored dots, some clustered together, some standing alone in a far corner of the town.

The dots, displayed in mere seconds, represent crime scenes. And whether the data represent a day, a month, or a year's worth of activity, police can tell at a glance how and when they should deploy their officers.

Stephen Sacks, professor of economics, at the E911 center of the UConn Police Department. The software Sacks has developed assists police departments in deciding how best to use their resources. Photo by Peter Morenus

"Police departments can spend months, years, refining their beats," says Sacks, a veteran economics professor at UConn who specializes in operations research. "With this system, they can redraw beats in minutes."

Sacks calls the software he developed PoliceDat. It can give police officers or planners a snapshot of recent crimes in any given district: what types of crimes have occurred there; what time of day they occurred; and how concentrated the crime is in any given area.

"It's very illustrative," says James Donnelly, director of public safety and telecommunications in the New Britain Police Department. "It shows you a pattern, and that can help you introduce a counter-meas ure - whether it's surveillance, flooding the area with officers, or maybe an undercover operation."

For Sacks, the work is clearly a labor of love. He smiles broadly as he moves from one aspect of his software to another. He proudly displays the certificate of merit he received from former Hartford Police Chief Joseph Croughwell, commending him for helping the city learn how to best deploy their patrol cars, and the New Britain Police Department arm badge that was presented to him.

And he clearly enjoys demonstrating what his systems are capable of.

Sacks has crafted or improved three basic programs: WinPCAM (a Windows version of a patrol car allocation model), which calculates the minimum number of patrol cars needed to meet various criteria; PoliceDat, which allows flexible and rapid display of police data, including where crimes have been committed by date, time of day, neighborhood, or distance from a local landmark; and the Desktop Hypercube, which allows planners to design patrol districts and evaluate them by displaying various performance measures.

With Hypercube, police departments that have entered all the pertinent crime and response data can measure their officers' workload - how much time each officer spends cruising his beat, how much time responding to incidents.

Sacks, breaking New Britain down into five segments, or beats, on his laptop, clicks on one and, instantly, the software shows how often that officer has responded to trouble on other beats. If he spends too much time in another area, Sacks says, planners can see whether - and how - they should juggle the coverage.

"You want this to be even, so all the areas of town get equal coverage. If the officer in one sector spends too much time helping out in another neighborhood, the response times in his own area will suffer," Sacks says.

Sacks, who joined UConn's economics department in 1971, says he began studying crime pattern analysis in the mid-1980s, leaving behind his earlier studies of market socialism.

Studying at the Massachusetts Institute of Technology during a sabbatical, he audited a course on urban operations research, and he was hooked. The MIT professor whose course he was taking had boxes of crime data from the Orlando (Fla.) Police Department that he had no time to work on. So Sacks turned his research interests "180 degrees," and left behind years of research on Yugoslavia.

"It was fascinating," Sacks says. "Here you had all these police departments all over the world, and many of them were not sure why they did what they did. It fit into my economics background, because these cities were wasting resources - patrol cars are scarce resources - by not properly deploying their fleet."

Sacks started his research in Orlando, using the data he obtained at MIT; developed the Hypercube; and went to work with Orlando police planners. By 1993, based on his efforts, Orlando had redrawn their patrol districts.

Today, Sacks is working with a variety of police departments in Connecticut, including New Britain, Bloomfield, and the Capitol Region Council of Governments. Police administrators in Vernon and West Hartford also are reviewing the software.

"There's a vast range of sophistication in police departments," he says. "Some are still using paper; some are pretty high tech but use only part of what's available; still others are using a lot of technology."

Sacks says that 10, or even five years ago, local police departments couldn't use his software: "The computers were too slow, there wasn't enough memory, it was too expensive. Now," he says, "we can do wonderful things."

Donnelly, the New Britain planner, says Sacks is one of just a handful of people in the country doing analysis of police operations.

"He's one of the few people who will roll up his sleeves and see how many officers to assign, where, and during what time of day," Donnelly says. "It's a neglected art."