Springboard’s annual Rise 2020 virtual conference concluded last week with a highly anticipated panel discussion on the role data science can play in leveraging social impact. DeRay McKesson, a civil rights activist and one of the leading voices of the Black Lives Matter movement, discussed the work of Campaign Zero, an organization he co-founded in 2015 that uses data-driven policy frameworks to end police violence in the United States.

At the time of writing, police have killed 839 people in the US in 2020 alone—a rate on par with previous years, despite much of the country going under lockdown for several months at the onset of the COVID-19 pandemic. The data paints a picture of a systemic problem with no easy solution: Black people are three times as likely to be killed by police as white people, and rates of police violence are not, in fact, correlated with levels of violent crime. 

This data matters now more than ever, as legislators and activists in at least 13 US cities, including Seattle, Austin, New York, and Los Angeles, are campaigning to defund their police departments and reallocate budgets to alternative forms of public safety. But these aren’t the cities most besieged by police brutality. Current data from MappingPoliceViolence.org identifies eight of the 100 largest police departments that kill Black men at a higher rate than the national homicide rate, including Reno, Oklahoma City, St. Louis City, and Scottsdale. 

“What’s wild to us is these aren’t cities you read about in the news, but these are the places where the problem is most acute,” said McKesson during his Rise 2020 session. “As organizers, if our solutions don’t hit these places, then it doesn’t matter.”

Using data to propose better policy solutions

No definitive dataset currently exists that charts all instances of police brutality, regardless of whether or not the officer was on-duty and if a weapon was involved. When it comes to using data to illustrate a social injustice, the rules are determined by whoever creates the database, which radically alters the way a problem is framed. 

For example, Fatal Force, a database maintained by The Washington Post, only includes killings that occurred while an officer was on-duty and in which a gun was involved. 

“This means that George Floyd is not on that database because he wasn’t killed by a gun,” said McKesson, referencing the high-profile incident in Minneapolis earlier this year where a white police officer, Derek Chauvin, knelt on the neck of George Floyd, a Black man, for eight minutes and 46 seconds. Floyd died shortly after.

Part of Campaign Zero’s work is to aggregate data from disparate sources and reconstitute it to aid in policy decisions. For example, requiring officers to wear body cameras and undergo implicit bias training have become go-to policy solutions for addressing police violence, but the data shows that these measures do little to reduce police misconduct. 

As for community representation, “the percentage of Black officers does matter, but not until the department is over 35% Black,” said McKesson, who hosts the social justice podcast, Pod Save the People.

Data on police violence exposes a flawed system 

While data has been used extensively to document the problem of police misconduct, it can also be used to examine an underlying justice system that condones the fatal use of force. Campaign Zero reviewed police union contracts in nearly 600 US cities and found that 84% of these contracts imposed at least one barrier to holding police accountable. 

In addition to creating a public database of police union contracts, McKesson also published the Law Enforcement Officer’s Bill of Rights so the general public could understand the existing provisions that are designed to protect police officers. These include preventing an officer’s name or picture from being publicly released or allowing officers to wait 48 hours or more before being interrogated after an incident. 

The protections granted by police unions explain why so few police officers who use excessive force on unarmed civilians are convicted of murder, often receiving lesser charges of manslaughter—or, as in the recent fatal police shooting of Breonna Taylor, “wanton endangerment.” 

“Some of the data we deal with is numbers and making the case. The other set of data we deal with is helping people understand their policies and how they stack up,” McKesson explained. “It is impossible to get the transformative change people want unless you engage the police union contract.” 

Using data to understand the state of policing involves so much more than just crunching numbers, says McKesson. “We are sensitive about understanding data not just as numbers on spreadsheets but people’s stories, their lived reality.”  

Using predictive analysis to compare outcomes

As state and local governments search for data-driven solutions to police violence, there is a greater need for predictive analytics to show the potential outcomes of criminal justice reforms like banning chokeholds, ending broken windows policing, and establishing civilian oversight of police investigations. “We try to help people think about what’s possible in their city by showing them case studies from cities that are similar to their own.”

One of Campaign Zero’s upcoming projects is a dashboard that allows cities to compare their outcomes with peer cities of similar political leanings and demographic makeup. Often, progressive cities like New York or Seattle are cast as “model” cities in the movement to end police violence, but every locality has its own approach to criminal justice reform informed by its population.  

McKesson stresses that it’s important to defer to the data, even if the recent explosion of Black Lives Matter protests across the country and public condemnation of police violence suggests things are changing for the better. “The data is important to us—we use it to guide us so we can figure out what the actual wins are because we worry sometimes that people think that the conversation changing is enough,” he said. 

“The police actually win in a context where talking about the problem differently is synonymous with [solving] the problem.” 

For more Rise 2020 coverage, check out our recap on how data science can influence election outcomes and tips on transforming your career in a post-pandemic world.