Building an Inclusive Public Analytics Platform for Police Data

How many police officers does your town have? This isn’t pub trivia so go ahead and Google it, I’ll wait.

Any luck?

If you live in one of the largest cities in the U.S. (NYC, LA, Chicago) you might have found the answer pretty easily. Anywhere else — probably not. And if your Google skills were good enough to find it — how do you know how many officers your city should have? In 2016 the median police department had 1.8 officers per 1,000 residents, but some cities have less than 1 officer per 1,000 people while others have over 4. There are over 18,000 law enforcement agencies in the U.S., but little information to understand how they compare to one another. I’ll give you some more policing facts later, but first let’s talk about why this matters.

The lack of publicly available, contextualized information on your local police results in a paradox of democratic accountability. All levels of government rely extensively on data analytics to carry out work on the public’s behalf. I should know; I’ve consulted on data analytics for many of them. But all too often this work is done without incorporating the perspectives of the whole community, which too often results in tragic consequences for low-income residents and people of color, as Virginia Eubanks and Cathy O’Neil have shown.

We can’t put the genie back in the bottle — when we ask the government to efficiently carry out complex tasks like providing public safety, analytics is a necessary tool. To maintain our government of the people, we require accountability by the people. And in order to scale government accountability to the complexity of our modern world, we need strong analytics.

If you’re reading this, you’re probably aware that analytics are having a moment. The other day I looked up something called VORP (an estimate of how much value a player contributes above the average available player) for my favorite basketball player, LeBron James. While sports fans can access the resources of a soon to be $2 billion sports analytics industry when making their fantasy player picks, when it comes to analytics on the people we trust to keep us safe, we’re often left sifting through PDFs.

But I’m changing that. I’m building an analytics platform to foster an inclusive discourse about local government services including policing. Don’t worry, I’m realistic. I don’t expect people to suddenly sit at a barstool and debate Officer Rios’ performance this year compared to Officer Thompson’s incredible 2011. But, we can make a lot of progress by starting from a shared knowledge of how many arrests for rape or homicide our town made last year — and then move forward together.

More on what’s next in a minute, but back to those police facts I promised:

  • In Albuquerque, the police department has 236 fewer officers than ten years ago, and, as the department’s size has declined, so has the arrest gap between black and white residents.
  • Atlanta has made some of the greatest gains in shrinking the black-white gap for drug arrests, but the gaps that remain are still staggering: in 2016, 1 black person per 100 was arrested for marijuana possession compared to than 1 white person per 1,000 arrested for the same crime.
  • And Madison, Wisconsin — where I came up with the idea for this project — has one of the largest black-white arrest rate gaps in the country with black people nearly 9 times more likely to be arrested for committing aggravated assault than white people.

These facts are not judgments or policy statements — they are the beginnings of conversations that need to happen. They are evidence of successes and challenges communities need to engage with, together. Importantly, they are not just points of data — they are information contextualized through comparison. Also, importantly, they are drawn from public sources and can be independently verified by all sides.

To demonstrate the possibility of an inclusive public analytics platform, I’ve created fact sheets on America’s 100 largest police departments. Each sheet relies on just a few key pieces of data from the FBI’s Uniform Crime Report and the U.S. Census to compare the size and arrest patterns of one city’s police to those of its neighbors and national peers.

Example chart from the Madison Fact Sheet showing the arrest rates for different offenses, by race, in 2016 compared to national and regional averages.

The fact sheets are a beginning — a demonstration of what we, as the public, can do to learn about our local government and increase its accountability to the people it serves. That’s why I’m reaching out to communities across the country to learn more about their local conversations around policing and to find out where analytics could be helpful in moving those conversations past conjecture and anecdote.

To support this outreach and continue to develop the platform for local government performance analytics, I’m partnering with RootProject to launch a crowdfunding campaign. I need your support to build a platform that is inclusive, open, and free. Learn more about how you can get involved here.