New York City Embraces Data-Driven Crime Fighting

Nowhere have declining crime rates been as dramatic as in New York City. In 2013, Manhattan had the lowest number of shootings, burglaries, and murders since the city began keeping formal records in the early 1960s. Crimes during 2014 were also historically low. Why is this happening? Experts point to a number of factors, including demographic trends, the proliferation of surveillance cameras, and increased incarceration rates. However, New York City would also argue it is because of its proactive crime prevention program along with district attorney and police force willingness to deploy information technology aggressively. Cyrus Vance, Jr., New York County's district attorney, is vigorously mining more crime-fighting information from the data collected by the city to drive crime rates even lower. He believes that New York could get crime rates to zero-if one looked harder at the data. There has been a revolution in the use of big data for retailing and sports (think baseball and Moneyball) as well as for police work. New York City has been at the forefront in using data intensively for crime fighting, and other cities have replicated its CompStat crime-mapping program. CompStat features a comprehensive, citywide database that records all reported crimes or complaints, arrests, and summonses in each of the city's 76 precincts, including their time and location. The CompStat system analyzes the data and produces a weekly report on crime complaint and arrest activity at the precinct, patrol borough, and citywide levels. CompStat data can be displayed on maps showing crime and arrest locations, crime hot spots, and other relevant information to help precinct commanders and NYPD's senior leadership quickly identify patterns and trends and develop a targeted strategy for fighting crime, such as dispatching more foot patrols to highcrime neighborhoods. Vance and his team think much more can be done with data to stop crime. While dealing with more than 105,000 cases per year in Manhattan, New York's assistant district attorneys did not have enough information to make fine-grained decisions about charges, bail, pleas, or sentences. They couldn't separate minor delinquents from serious offenders quickly. In 2010, Vance's team created a Crime Strategies Unit (CSU) to identify and address crime issues and target priority offenders for aggressive prosecution. Rather than information being left on thousands of legal pads in the offices of hundreds of assistant district attorneys, CSU gathers and maps crime data for Manhattan's 22 precincts to depict criminal activity visually, based on multiple identifiers such as gang affiliation and type of crime. Police commanders supply a list of each precinct's 25 worst offenders, which is added to a searchable database that now includes more than 9000 chronic offenders. A large percentage are recidivists who have been repeatedly convicted of grand larceny, active gang members, and other priority targets. These are the people law enforcement wants to know about if they are arrested. This database is used for an Arrest Alert System. When someone considered a priority defendant is picked up (even on a minor charge or parole violation) or arrested in another borough of the city, any interested prosecutor, parole officer, or police intelligence officer is automatically sent a detailed email. The system can use the database to send arrest alerts for a particular defendant, a particular gang, or a particular neighborhood or housing project, and the database can be sorted to highlight patterns of crime ranging from bicycle theft to homicide. The alert system helps assistant district attorneys ensure that charging decisions, bail applications, and sentencing recommendations address that defendant's impact on criminal activity in the community. The information CSU gathers and disseminats through the arrest alert system differentiates among those for whom incarceration is an imperative from a community-safety standpoint and those defendants for whom alternatives to incarceration are appropriate and will not negatively affect overall community safety. If someone leaves a gang, goes to prison for a long time, moves out of the city or New York state, or dies, the data in the Arrest Alert System are edited accordingly. In speeches praising intelligence-driven prosecution, Vance often cites the example of a 270- pound scam artist who for over a decade made a living by bumping into pedestrians in the Times Square area and demanding money, claiming they had broken his glasses. He had been convicted 19 times but only for a misdemeanor charge and never served more than 5 months in jail. When the CSU flagged him after his arrest in July 2010, he was convicted of felony robbery and sentenced to 3½ to 7 years in prison. Information the CSU developed helped Vance's Violent Criminal Enterprises Unit break up the most violent of Manhattan's 30 gangs. Since 2011, 17 gangs have been dismantled. According to New York's chief assistant district attorney, Karen Friedman Agnifilo, murders dropped from 70 in 2010 to 29 in 2013 because the DA's office and police now had the information to identify the people driving crime in Manhattan and to move these people off the streets and behind bars. There's another side to this story, however. When prosecutors begin to compile databases for data-driven crime fighting, one needs to ask what people have been selected for inclusion in these databases, what are the selection criteria, and how harmful is this practice? Could the criminal justice databases include people who really shouldn't be there and nevertheless are targets for police scrutiny? According to Steven Zeidman, director of the criminal-defense clinic at the City University of New York (CUNY) School of Law, the answer is yes. More than one thousand people are arrested in New York City each day. An overwhelming and disproportionate number are people of color arrested for minor offenses such as riding a bicycle on the sidewalk or jaywalking. Zeidman recalled a time when he was in court with a teenager arrested for jaywalking. The arresting officer said he had stopped the young man because he was wearing a red shirt that was known to be a gang color. The young man was not a gang member, but he's probably in the database.

Case Study Question

1. What are the benefits of data-driven prosecution for crime fighters and the general public?

2. What problems does this approach to crime fighting pose?

3. What people, organization, and technology issues should be considered when setting up information systems for intelligence-driven prosecution?