How sharing police data can improve relationships with communities

Panelists at an event at Harvard's Kennedy School argued that data on officer-involved shootings and body cameras could help improve citizens relationships with the police. But the technology should be used in conjunction with a larger conversation on race, class, and policing, they said.

Many people trace the beginning of efforts to use data and statistics to prevent crime to the early 1990s, when the New York City police department began using maps and statistics to track where crime occurred most frequently. More recently, departments have begun releasing crime data to the public amid a growing national outcry about shootings and uses of force by police directed at primarily young black men.

But Nicole Wong, who served as the White House’s deputy chief technology officer until 2014, argues that the complex relationship between race, crime data, and policing instead dates back a century earlier — to the pioneering efforts of African American journalist and activist Ida B. Wells to discover the true causes of an epidemic of lynching sweeping the American South.

“I was doing some of my own research, trying to find women in data science, and I came across one of our earliest data scientists — Ida B. Wells,” Ms. Wong said during a panel discussion focused on criminal justice and data analysis at Harvard University’s Kennedy School on Friday.

In 1889, Ms. Wells began examining a series of reports in the Chicago Tribune on lynching, trying to distinguish patterns that linked the lynchings together.

After examining 10 years of data — Wells reached what was at the time a surprising conclusion — contrary to popular perceptions used to justify lynching — 50 percent of the black men and some women were lynched for crimes such as “having a bad reputation,” or “writing an insulting letter” or for no cause at all, Wong told the panel.

Wells’ use of official data in her reporting helped dispel widespread myths that lynching occurred mainly as a punishment for rape and murder, she added.

“For me, that was a lesson in [the idea] that we certainly want to use data to improve our criminal justice system, but how we think about that data, and how we use that data is vitally important,” said Wong, who previously worked as a lawyer at Twitter and Google before coming to the White House.

But more than a century after Wells’ expose on lynching, many police departments are only beginning to make crime data freely available to the public. Why now? It's a bid to improve long-fractured relationships with minority communities as the country engages in a national dialogue about race and policing, other panelists said.

“The lack of that data had a significant effect on how people viewed the law enforcement system in this country, how do you hold that system accountable if you don’t know what it’s doing?” said Clarence Wardell, a Presidential Innovation Fellow and affiliate of Harvard’s Berkman Center for Internet and Society.

Earlier this year, he began working on the Police Data Initiative, a White House program formed in May in the wake of the shooting of unarmed teen Michael Brown by a police officer in Ferguson, Mo. The program aims to provide data particularly on officer-involved shootings across the the country, as means to help build trust between the police and citizens increasingly alarmed by policing practices such as the controversial “stop and frisk” policy used for many years by departments across the country.

“For a lot of departments, it’s not like they were actively against [releasing data], they just didn’t really know how to do it,” Dr. Wardell added. “It’s great to release reports, but let’s see what’s underneath, so [citizens] can decide for themselves.”

The White House program has also benefited departments internally, he said, noting that many police departments have struggled to build “early warning systems” to predict officers that might be involved in shootings or be likely to use excessive force.

But for police, how best to employ technologies like body cameras is still an ongoing debate, with some departments crediting it with dramatically decreasing the use their use of force, as others debate its ethics and how best to handle the reams of data the cameras produce.

“In policing, our own internal bureaucracy can be challenging,” said Lt. Dan Wagner, lead crime analyst at the Cambridge Police Department, who is currently on a leave of absence to pursue a graduate degree at the Kennedy School. “Cops aren’t gonna want to wear body cameras if they know some supervisor’s going to jam them up by looking at the video and saying ‘Hey, you did some little thing wrong,’ and using it as a means to administer internal punishment."

But, he added, “If it’s to scrutinize the use of force, I think that’s certainly appropriate.”

Cambridge partnered with Massachusetts Institute of Technology researcher Cynthia Rudin, who was also part of Friday’s panel, to use machine learning technology to help identify and solve burglaries. In that city, the police are still considering whether to adopt body cameras, Wagner said.

The department has been cautious about other technologies that are touted as being able to “predict” crimes, such as efforts to wholly “vacuum up” residents’ social media information to determine whether a crime is likely to occur. After learning about one such effort in Florida – where the information was held indefinitely – Wagner said, he balked.

“I couldn’t fathom the people of Cambridge being okay with that,” he said, drawing laughter from the crowd.

Professor Rudin, who teaches at MIT’s Sloan School of Management, said a key obstacle for predictive policing technology such as her “Series Finder” tool is getting access to policing data. While she was able to partner with police in New York and made her code publicly available, the department didn’t want to give her access to their data. The department instead ran her software itself.

But Wong, the former chief technology officer, who is now affiliated with Harvard’s Berkman Center, said data should be used cautiously.

“I really do believe in the power of data, but I also believe it is never neutral,” she said.

Several panelists touched on the role of the media to educate citizens by using policing data. But they said media coverage often focused on a single incident that “looks good on TV,” as Wong described it.

“We might focus on one very profound incident that we all see, but we should also be aware of the 278 other shootings that happen,” she said.

Wagner, the police officer, said the focus on specific shooting incidents sometimes obscured a larger conversation about the relationship between crime, race, and social class in America, which he said wasn’t always discussed.

“There are problems in policing, but I think that the problem is not the young black or brown man and the cop on the street, “ he said. “I think there are larger, more complicated, social, economic, cultural issues in our country that that is one symptom of, and it’s a serious symptom.”