Lessons learned from a year of opening police data

by and
President Barack Obama speaks to the press after a meeting with members of the President’s Task Force on 21st Century Policing. (Photo credit: Chuck Kennedy/Official White House Photo)

It’s been almost a year since the White House first announced that it would be leading an effort to unite law enforcement agencies around the goal of achieving greater transparency through data. In April, the White House Police Data Initiative (PDI) celebrated its progress by gathering leaders in the field for a two-day event to discuss the challenges and successes of releasing open police data to the public.

The initiative began with 21 participating jurisdictions last May. Since then, that number has more than doubled to 53 jurisdictions that have published over 90 datasets in the process. In light of commitments by 32 additional agencies and organizations, Sunlight reaffirmed its dedication to the ideals of the initiative by pledging to add all datasets opened by participating agencies to Hall of Justice, a repository of criminal justice information launched in February.

As partners of the initiative since the beginning, Sunlight was excited to participate in these important conversations and help represent the voice of civil society at the event. Open data leaders in law enforcement advocated strongly for other agencies to get on board with proactive transparency practices, but the subsequent breakout sessions saw stakeholders delving into deeper analyses of how this could be accomplished.

The privacy implications of personally-identifiable information

The first day’s event opened the table to advocates for victims of domestic violence, sexual assault and stalking, many of whom had strong opinions about the dangers of providing public access to personally identifiable information online.

According to a survey conducted last year by the National Domestic Violence hotline, 60 percent of women who had not interacted with police following an incident avoided law enforcement out of a desire for privacy. Some feared the inclusion of addresses in incident-report data could subject victims to pursuit by their abuser, discouraging them from reporting other crimes committed against themselves or others. FTC Chief Technologist Lorrie Cranor spoke about the dangers of re-identification when opening incident-level data, noting that the disclosure of demographic information in small populations can run the risk of spotlighting potentially sensitive details about individuals.

Refer to our series of briefs about what kind of data is publicly accessible in the criminal justice world to learn more about anonymization techniques and current practices in how governments are opening microdata.

Challenges in building community trust and developing standards

A key part of the conversation at day two of the event focused on the need for standardization of law enforcement data for open data practices to move forward efficiently.

Many in the law enforcement community noted that use of force, for example, may be defined in several different ways depending on the jurisdiction. Some shared that in their department, an officer raising a weapon could be considered a use of force incident, while others noted that threatening language might meet this definition. Such discrepancies in what data could look like make comparisons difficult and benchmarking progress nearly impossible.

Read our blog post about the importance of criminal justice data standards for more examples on how varied terms and collection techniques can be.

The importance of storytelling and context

Leaders in police departments from around the country emphasized the need for telling a story about their open data. Los Angeles County Sheriff’s Department Chief Data Officer Wendy Harn shared her experience with making open data available to the public: “I thought a daily data dump was being transparent, it really isn’t. It’s just saying, ‘Here’s my data, figure it out.’”

She continued to emphasize that departments must work to interpret their data in a way that provides context for the public. For example, the county sheriff’s site includes both incident-level data on deputy-involved shootings as well as a visualization illustrating aggregate counts of the types of incidents that were logged over the last five years. To provide context, the site also gives in-depth explanations of what kind of incident falls into each descriptive category.

When police departments are at the forefront of revealing what their data means, it’s less likely that the media or public will misconstrue it in ways that paint the wrong picture. Such misunderstandings can damage a police force’s reputation in ways that lead to community distrust and diminished confidence in the department’s ability to protect the public.

Painting a picture of what data looks like includes disclosing why data was collected, how it was collected, when it was collected and what this data means for the department and the community as a whole. When data the public is asking for cannot be released, departments should convey their reasoning, which usually relates to privacy concerns like victim identity protection.

During one of the working sessions, “Open Data, Accountability, and Privacy,” some members of the law enforcement community noted that sometimes privacy laws prevent departments from providing context for their data. Investigatory records may be exempt from public records laws, for instance, occasionally causing valuable bits of information to be excluded from telling the whole story of the data.

Ideally, departments will also find ways for the public to easily access and visualize open data on user-friendly portals. Public engagement with the data in combination with contextualization from police departments creates transparency that ultimately leads to a mutual trust that will benefit communities moving forward.

Return on investment

Police chiefs spent time dispelling myths surrounding the resources it takes to open up data as well. Panelists from several departments involved in PDI reflected on the fact that no additional cost to departments was incurred in the process, and that opening datasets actually saved staff resources and reduced time spent responding to requests for information. As Dallas Police Chief David Brown put it, the trust his city’s open data program generated within his community has made every effort entirely worthwhile. “The ROI on open police data is high and will pay dividends for years to come,” Brown said.

The next year in open police data

Over the next year, it’s expected that PDI membership will continue to grow rapidly, with more and more police departments across the country making commitments to release high-value datasets to the public. We will be updating Hall of Justice with these developments to help provide context to the prevalence of criminal justice data and open police data around the country.

We also encourage departments involved in the program to focus on building community engagement efforts to interact with the datasets and provide input on datasets for release. We hope that law enforcement will further these efforts by continuing to gauge public demand for high-value datasets through attention to the volume of FOIA requests from citizens and targeted outreach to communities.

The Police Data Initiative highlights these efforts in cities across the country; this includes a special preview of upcoming police datasets for young coders in New Orleans and a hackathon in Indianapolis centered on citizen complaints and use-of-force data that the police department worked together with Code for America to release.

While hackathons are a great way to engage community members, they can also be isolating in that they don’t necessarily attract community members without certain technical skills. Police departments should think outside the box to find more inclusive methods for engaging the public. As such, these efforts must be broadened to include regular, proactive communication with community and advocacy groups. For instance, the city of Orlando convened sexual assault and domestic violence victim advocates around addressing privacy issues with open data, leading to a larger discussion at the first PDI event we mentioned earlier in this post.

We encourage police departments to share with us via our OpenGov Voices blog any innovative approaches they have taken to community engagement, and look forward to following the conversations that develop out of themes participants explored at the event.