One Man Built a 30,000-Record Police Death Database Because the Government Wouldn't

There is no complete federal database tracking how many people die during encounters with police in America. Not a mandatory one, not a comprehensive one. That gap in federal oversight is what prompted Fatal Encounters, a volunteer research project that has spent years doing the job no federal agency was ever required to do.

The project, founded by journalist D. Brian Burghart, has compiled more than 30,000 records of deaths that occurred when police were present or that police caused. That includes on-duty and off-duty incidents, criminal acts, line-of-duty deaths, accidents, and deaths at the local, state, and federal level. Burghart's team says it completed a systematic state-by-state, year-by-year search of the country by November 3, 2017, and has continued updating the database multiple times a week since.

For years, the working assumption pushed by government sources was that fewer than 400 people were killed annually during police actions. Burghart's data collection helped demonstrate that number was way off. Nobody had to speculate about whether the government was hiding something sinister. They just had to notice that nobody was counting seriously in the first place, and a private citizen with a spreadsheet did what taxpayer-funded agencies didn't.

Local, state, and federal agencies are not required to collect complete data on the numbers and characteristics of people killed by police. That's according to Burghart's own team, who point out this isn't a partisan grievance, it's a structural gap. If you want better data, the fix isn't outrage on social media. It's contacting your local, state, and federal representatives and asking why this still isn't mandatory in 2026.

What the database actually is, and isn't

Burghart has been explicit that Fatal Encounters is a research tool, not a running commentary or advocacy blog. In a message dated September 18, 2020, the project noted that "not much about the Fatal Encounters project happens on the website" itself. It's primarily a delivery mechanism for a Google Spreadsheet, with a searchable interface for people who want to filter down to specific categories.

Different research teams use the raw data for wildly different purposes. One team, according to the project, focuses only on police-related suicides. Another studies teenagers who died during police pursuits. The philosophy behind the project is that the underlying facts should be public and neutral, and it's up to individual researchers, journalists, and advocates to draw conclusions from it.

Burghart has also been upfront about the dataset's limits. In a June 2020 statement, he cautioned against treating the 30,000-plus records as comprehensive. "It's not," he wrote, adding that the dataset was never intended to be a finished product, but rather a first draft meant to identify names, demographics, dates, and locations, and to flag categories that need deeper investigation.

The project has also flagged planned upgrades, including new "threat assessment" columns covering whether a subject was armed or unarmed, what weapon was alleged, physical movement during the encounter, and whether the person was fleeing. Those categories matter enormously for anyone trying to have an honest conversation about use-of-force policy, because raw death counts without context about threat level tell you far less than people assume.

The academic partnership

Burghart's team partnered with a multidisciplinary group at the University of Southern California to produce what they call the National Officer-Involved Homicide Database, available for download after signing a data-use agreement. That partnership represents an attempt to move the project from a solo journalist's spreadsheet toward something with institutional academic backing, complete with a codebook explaining how records are categorized.

No single, mandatory, standardized federal system requires every police department in the country to report every death connected to an encounter with law enforcement. Congress could close that gap tomorrow if it wanted to. It hasn't.

The unresolved question isn't whether police kill people in numbers higher than old federal estimates suggested. Burghart's work already reshaped that public understanding. The unresolved question is why, years after that number was publicly debunked, there's still no legal requirement forcing every agency in the country to report complete, standardized data to the public that pays their salaries.