“In preventing terrorism, as in preventing crime, the secret is data, the intelligence that you make out of it and how you use that intelligence,” Bill Bratton told the audience at the CEB TowerGroup Financial Services Technology Conference in Boston. Bratton, the former head of the Boston, New York, and Los Angeles police departments, was invited to speak at the conference about “data-driven policing.” Also called “predictive policing,” today it is a widely-discussed big data example, building on a new approach to policing which Bratton has played a leading role in its development and implementation over the last two decades. In the aftermath of the Boston Marathon terror attack, Bratton’s talk was highly instructive. Here’s what I learned about the link between hometown and homeland security and the use of data to combat crime and terror.
As Bratton tells it, the revolution that happened in the early 1990s and subsequently led to a dramatic reduction in crime rates, was the shift to community policing. The key to this shift was a radical change in the philosophy of policing: Instead of focusing on response to crime, the police had started to focus on crime prevention. In focusing on prevention, the police had gained a new understanding of the causes of crime. Pointing out that crime rates have not increased with persistent high unemployment in recent years, Bratton described how with community policing the focus has shifted from general social issues traditionally blamed for crime such as poverty, racism, and unemployment, to the people who commit crimes—criminals and people committing crime of passion—and to controlling their behavior in a manner that is constitutionally correct, compassionate, and consistent. This led to a new focus on minor crimes, instead of the previous almost exclusive focus on major crimes. During his time as the chief of the transit police in New York City, “we were starting to focus on what the data was telling us,” Bratton said. “A relatively small number of people were creating an atmosphere of fear where other people found opportunity to commit more serious crime.”
Listening to the data was certainly an important part of the community policing revolution. But it did not start with technology. It started with paper maps hung on walls and push-pins indicating where crime has occurred. This pre-computer “data visualization,” helped Bratton and his colleagues make sense of “huge amount of information,” identify hot spots, and “put cops on the dots.” And there was another innovation not related to technology: Collaboration and sharing of information. When Bratton started to run the NYPD, they brought “all the precincts’ commanders together to have them talk about crime, why is it up or down, what’s working, what’s not.”
Desiring to enlist technology in support of this new philosophy and practice of policing, Bratton and his colleagues went to “the room full of computers” at police headquarters and asked for help in mapping the crime in the city. They were told, “It’s going to take six months, probably a year.” So Jack Maple, Bratton’s Deputy Commissionaire, went to Radio Shack and bought a $98 Tandy computer. This was the genesis of CompStat, for computer statistics, still in use today in New York and other police departments. While certainly a great example of a smart use of computer technology for data collection and analysis, it was clear from Bratton’s talk that CompStat is “a management philosophy or organizational management tool for police departments,” as the Wikipedia article puts it. Bratton kept enumerating in his speech the four elements of this management philosophy: timely, accurate, intelligence; rapid response; effective tactics; and relentless follow-up. “CompStat was about having a goal,” he said, “it was about focusing on the results we wanted to achieve.”
The new philosophy eventually made an impact beyond local police departments. Before—and even after 9/11—“just like the cops were focusing on response to crime,” Bratton told me after his talk, “Federal agencies were focusing on the external threat coming from off-shore. “ But this has started to change after 9/11 and Bratton sees the parallels with the new tenets of local police work: “We were beginning to develop a seamless web of timely, accurate, intelligence. We were beginning to develop collaborative response to threats and effective tactics. And if there were to be an event, and there were several, relentless follow-up on how it had happened. With the notable exception of the successful terror attack last week, there have been few because we have gotten better in the use of technology. In the age of big data, the more information we have the better the analysis.”
Homeland security became hometown security. According to Bratton, 75% of the terror threats that have been detected since 9/11, have been detected by local police. He recounts examples from his tenure as the head of the Los Angeles police department between 2002 and 2009 where police investigating robberies or stolen vehicles found connections to terrorists because they now were trained to look for these types of clues. Bratton himself in that period spent 40% of his time on terrorism, compared to the 2% he spent when he was head of the NYPD, a position he was appointed to a year after the 1993 World Trade Center bombing.
Bratton thinks a lot of progress has been made since 9/11, but “there is still weakness in the collaboration. We’re doing a significantly improved job since 9/11 of gathering the data and analyzing it. But because of the significant number of agencies involved, it’s the sharing amongst them that’s still the issue.” And not just sharing of data, but also sharing of analysis: “If your analysis is different from mine and we are not sharing these differences, then we are missing opportunities,” says Bratton.
Listening to Bratton, I thought that the lessons to be learned from the success of community policing can be summarized in one word, people, both the people trying to prevent terror and the people planning terror attacks.
Big data, more data, is not going to help on its own. More seamless sharing of more data is important but will turn out to be a futile investment if the people responsible for hometown and homeland security do not share their intelligence. Even more important, if they don’t have, as CompStat had, a clearly identified goal, an understanding of what to measure and analyze.
So another lesson is the need to question traditional assumptions about the causes of terrorism and the people involved. Is “religion” in the prevention and understanding of terror the equivalent of “unemployment” in the prevention and understanding of crime? Is there a possible connection, related to what motivates people in modern societies—and drive some to violence against others or against themselves—between mental illness, mass shootings, and acts of terror, as sociologist Liah Greenfeld (full disclosure: my wife) asks in her book, Mind, Modernity, Madness? Is there anything being done to prevent a terror attack that may be perpetrated by an individual with a similar psychological profile but with no adherence to any “religious fundamentalism”?
It has been a recurring theme of this blog that more data is not better than people, that correlations cannot substitute for causation. In preventing terrorism, as in preventing crime, a better understanding of causes may help in the efficient collection of data and its intelligent use.
[Originally published on Forbes.com]