In the government’s hands, data is powerful. Federal, state, and local governments gather massive amounts of data on Americans in the name of public safety, national security, and welfare-fraud prevention. Funneled through inscrutable algorithms, that trove of information can cut people off from loans, housing, and jobs.

But data-gathering isn’t just the government’s game. Since recording technology became portable and cheap, citizens have been keeping tabs on their governments, too. Steve Mann, a wearable-computing pioneer, called it “sousveillance,” switching out the “sur” in surveillance—French for “above”—into “sous,” for “below.” When a bystander shot a grainy video of Rodney King getting beaten by a group of police in March of 1991, that was one of the earliest examples of sousveillance.

Brandon Anderson wants to systematize sousveillance, by creating a platform for organizing citizen-supplied information about police.

Anderson grew up in Oklahoma, where his relationship with law enforcement was characterized by tragedy. His mother spent much of his childhood in prison, and his longtime partner was killed by police in Oklahoma City in 2007, when Anderson was a satellite engineer in the Army and had just returned from a tour in Iraq. (He wrote about that experience in  a moving personal essay on Medium.)

“Police terror has, for me, been a constant in my life and everything that I do,” Anderson told an audience at Georgetown University last week, where he was presenting at a conference on surveillance and communities of color.

After nearly five years in the Army, Anderson worked a variety of jobs and eventually matriculated at a community college in the Seattle area. Two years later, he transferred to Georgetown. And since he graduated last year, Anderson has been working on creating an app to allow communities to gather and share information and feedback about their police forces. He’s especially targeting low-income communities and communities of color—the most heavily policed groups in the country.

Anderson’s app, which is still under development, is currently called SWAT—Safety with Accountability and Transparency—but he says he’s thinking of changing the name. Anderson drew inspiration for his platform from Waze, a popular driving app that allows users to mark obstacles like potholes for others to see and avoid. Waze passes along that user-generated information to local governments to prod them into filling the potholes—but until that happens, it allows drivers to avoid them.

When he initially set out to build the app, he was looking for a way to automate the official police-complaints process, hoping to encourage more people to submit their grievances. But after he polled potential users, he found that a deep-seated mistrust of official reporting channels was keeping people from submitting the reports. Many had lost faith in the system after repeated complaints didn’t make a difference.

So, instead, he’s creating a separate, unofficial data-gathering operation—“a mapping system of police performance”—which he hopes will gain community members’ confidence and win their engagement. “We’re all on board with creating trust in community government, but we’re not going to wait around for them,” Anderson said. “We’re gonna build this shit ourselves.”

The app will allow users to record experiences with police—Anderson calls them “micro-moments”— that can be placed on a map and shared with nearby community members. Others can choose to support the user that reported a moment, or take action based on the user’s report. Over time, the platform will aggregate data about the sorts of events that characterize each community’s run-ins with the police, information that Anderson hopes can help governments and police forces make adjustments to their policies and practices.

When he took the stage at Georgetown last week, Anderson began by recounting a powerful string of stories about his family’s and partner’s destructive interactions with police in Oklahoma. I asked him if he would put a similar focus on narrative storytelling in his platform. “People understand data better when it’s humanizing, when it talks from a point of view—that, every person can understand,” he said. “Data is not felt. I want to encourage people to feel these stories, not to think this data.”

That’s where community data-gathering and government surveillance diverge. When the state gathers information about people, it’s usually not in prose. The personal data that the government has access to is made up of facts and figures: a missed loan payment, an arrest record, a list of IP addresses accessed on a mobile device.

But numbers alone aren’t enough to galvanize a community. Sousveillance thrives on shaky videos, eyewitness photos, and moving stories, amplified by social networks, both online and offline.

Right now, it can be hard to know what to do with an incriminating photo or a story about a negative interaction with law enforcement. People with a strong following on Twitter or Facebook might be able to broadcast their news widely, but the never-ending sea of social content can swallow others’ stories. If there’s a central destination for depositing data about police, it can become a valuable vehicle for keeping government accountable.