Brilliant Products

A Reporter Developed An Algorithm That's Spotting Serial Killers

Big data has proved incredibly useful for making predictions and solving problems in everything from politics to baseball. Starting in 2010, Thomas Hargrove began bringing the power of big data to law enforcement: he developed an algorithm that pulls stats from crime databases in the hopes of spotting trends that could point to the work of serial killers.

FBI On You

In 2004, Thomas Hargrove, then a news reporter, was researching statistics about prostitution. When he received the FBI's Uniform Crime Report (UCR) from the database library so he could check the numbers, he also happened to receive the Supplementary Homicide Report: a list of every single murder reported to the Bureau that year. As you might guess from his original assignment, Hargrove was his paper's go-to numbers guy. His affinity for stats piqued when he looked through the heavily detailed murder report, and it occurred to him that an algorithm might be able to spot trends in the data. "I don't know where these thoughts come from," Hargrove told Bloomberg, "but the second I saw that thing, I asked myself, 'Do you suppose it's possible to teach a computer how to spot serial killers?'"

He got help from a University of Missouri grad student to put all of the data into statistics software, then spent months developing an algorithm that could find commonalities among unsolved cases. To do that, he used a solved case: the serial killer Gary Ridgway, a.k.a. the Green River Killer. After performing hundreds of trial and error tests that failed to identify Ridgway's real victims, Hargrove narrowed his formula down to four distinct trends: location, gender, age range, and how the victim was murdered. The algorithm identified Gary Ridgeway right away. It also identified many other caught serial killers, while also flagging a handful of unsolved murders that could be connected.

One of the first serial killers that Hargrove potentially identified that hadn't been caught yet was in Gary, Indiana. He notified the department of what he'd found, sending both emails and snail mail, but he never heard back. Later, a serial killer was caught, and the man said he'd killed many others since the 1990s. Hargrove couldn't speak to anyone involved in the case once the trials were underway, but Hargrove suspects that the killer was the same man he notified the police about. It was presumed a classic case of avoiding innovation in favor of tradition of which police departments around the country have been accused for decades.

The Billy Beane of Brutal

One of the issues with identifying serial killers using Hargrove's algorithm is the pure lack of data. Not all police departments have access to each other's unsolved murders and not all report them to the FBI, so compiling all of it would be an enormous task—and that's not to mention adding new information as it comes in. Still, Hargrove is determined to change the criminal justice system at his own level by using all available information, publishing them on The Murder Accountability Project—an organization he founded in 2015— and crowdsourcing the effort to spot trends. Whether it's someone mourning a victim and seeking justice or a numbers-savvy investigator with a lust for justice, anyone can participate in making the world a safer place.

Watch And Learn: Our Favorite Content About Cold Cases

Thomas Hargrove On San Francisco's Unsolved Murders

From Cold Cases to Positive IDs: A Forensic Art Class Helps Uncover Victims

Written by Mike Epifani June 12, 2017

Curiosity uses cookies to improve site performance, for analytics and for advertising. By continuing to use our site, you accept our use of cookies, our Privacy Policy and Terms of Use.