Digital Crime Fighter
How to Increase the Reliability of Your IT Infrastructure Using Predictive Analytics REGISTER >
When William Bratton was sworn in as New York City's police commissioner in 1994 at the age of 47, he made what many considered an unrealistically bold promise: The NYPD would fight crime in every boroughand win. Crime would go down by 10 percent the first year and by another 15 percent the next. It seemed foolhardy: Everyone knew you couldn't win a war on crime. But Bratton delivered. Over 27 months, a computerized crime-mapping system he developed with deputy commissioner Jack Maple and police computer whiz John Yohe helped him to cut the number of shootings by 16.4 percent, homicides by more than half and serious crime in New York City by 33 percent.
Bratton landed on the cover of Time for his effortsbut found himself in hot water with New York Mayor Rudy Giuliani over who should take credit for the program. Bratton ended up getting ousted in the spat, but his Compstat crime-fighting technology lives on. About one-third of the nation's police departments, from Philadelphia to New Orleans, are now using it or are in the process of installing some form of Compstat to map crime and reform crime fighting. This spring, Giuliani remarked that he'd like to see the system applied to solving problems in city schools; school board members are seriously mulling a proposal.
Bratton, meanwhile, now a private police consultant for New York City-based Kroll Associates, is working with officials in Caracas, Venezuela, to apply Compstat there to ease a 30 percent rise in crime. "Bratton's Compstat program singlehandedly used technology to change the face of crime fighting and the effectiveness of not just one police department, but hundreds," says Eli Silverman, a professor at the City University of New York's John Jay College of Criminal Justice.
Not bad for what began as a doctored-up spreadsheet and a bit of hand-written code hammered out on an old 486 computer during a blizzard one weekend seven years ago. Yohe remembers taking the first stab at writing the Compstat program this way while stranded in the city one weekend in February 1994, after the Long Island Railroad shut down during a snowstorm. "Maple, in those first weeks of Bratton's term, would walk around asking people how many crimes were occurring," Yohe recalls, "but nobody could furnish that information."
Numbers that did exist were six to 12 months old, and not at all complete or detailedjust enough to fulfill the basic requirements of a federal crime statistics law. So Yohe began tinkering away at a program. "Until Bratton, people were paying lip service to crime fighting but nobody was doing it," Yohe says. "I found Bratton and Maple very inspiring."
The resulting Compstat technology is part data-mining and part spreadsheet software technology that enables crime fighters to find patterns, say, in the murder rate of a particular neighborhoodby time of day, location, weapons used and dozens of other variables. Commanders can then better identify and attack the problem. "Mapping technology can make you smarter about everything," Yohe says, from crime to garbage collection, depending on how you render the program.
Before Compstat, for example, Bratton says, "people didn't have a clue what was really happening" on the streets of New York. One precinct commander, when asked about the number of armed assaults in his neighborhood, repeatedly explained it away as a problem occurring around bars at night, Yohe recalls. But once Compstat was used to size up his assault problem, officials discovered that the altercations were occurring between 3 p.m. and 5 p.m., and that the "perps" were teenagers between the ages of 12 and 15 on the way home from school"not adults getting drunk at night outside a bar," Yohe says. "Before, we'd have to take the commander's word for it. Now we had a way to measure accountability. It was revolutionary."
Change, though, wasn't easy. Bratton says he found out early that it's not enough to have a champion for change at the top: You also need to force the career bureaucrats in the middle to accept it, tooor risk torpedoing even the most powerful of intentions. At first, resistance was so high to Compstat that in its first year, 75 percent of the city's precinct commanders had to be replaced, for either failing to work with the system or ignoring what Compstat analyses showed about crime in individual precincts, Bratton says. "If, week after week at the Compstat meetings, we found precinct commanders not performing to standards, we had to find someone else to do the job," he says.
How much opposition was there? Bratton and his reformers initially had to appeal to a police benevolent fund to get the cash needed to purchase Compstat's first computer maps and hardware. "The MIS department simply wrote us off," Yohe recalls. "You name it, they refused to give it to usthe hardware, the software, the training, facilities, equipment of any sorteverything."
To set up the NYPD's first LAN as part of Compstat, for example, Yohe literally had to run wires for it out a 13th floor window and back through a window on the eighth floor. "The department told us it would take a year to run the wires we needed, so in the meantime, I opened a window and threw [down the] wire that was needed," he said. "You weren't supposed to do this stuff, but we didn't care. For years people thought we weren't doing anything to fight crime and here we were, actually trying to turn things around. It was exhilarating."
But Yohe also felt the heat of the early changes. Many officers, at first, complained loudly that they didn't have the time needed to input crime data each week into the Compstat system, as required. Many precincts were late sending the data, and it was Yohe's job to get them in line. At one point, Yohe's photograph was pinned to a bulletin board in a Brooklyn precinct office, with darts crowded around his forehead. "I was hard on that precinct because they kept missing their deadlines," Yohe recalls. But Yohe persisted, pointing out that it took only 17 minutes to fill out the forms, at most. "I'd tell them that Compstat is not hard. Hard is being 75 years old and getting your brains beat out when you're crossing the street for a quart of milk, or being a single mom and having your kids sleep in a bathtub to avoid getting hit at night by stray bullets."
For Bratton, though, Compstat was chiefly about doing the public's business better, in much the same way a business or corporation seeks to do a better job achieving its goals. Says Bratton: "Every McDonald's knows how many hamburgers it sold yesterday, so it knows how many it needs to stock up for today. Similarly, by using Compstat, every precinct should know how many 911 calls there were on Friday night and, therefore, how many police cars are needed. It's all about mapping out the problem so you can bring about results."
Does Bratton ever reflect on those early days of Compstat and its legacy? In his book, Turnaround (Random House, 1998), he describes his battle to bring about change and remains adamant about the need to boost accountability of systems and personnel through the use of technology. "You can't hold anyone's feet to the fire if you have no way of measuring what they do," he says.
But others say Compstat was more than a story about results. "It's about leadership," says Silverman. "Bratton dared, and got others to dareand changed the world in ways that keep influencing others."
Donna Tapellini is a New York City-based technology and science writer whose work has appeared in a variety of print and online publications. Comments on this story can be sent to firstname.lastname@example.org.