An Algorithm That Catches Serial Killers Could Connect 200k Unsolved Murders

Wednesday, 07 March 2018 - 11:21AM
Technology
Wednesday, 07 March 2018 - 11:21AM
An Algorithm That Catches Serial Killers Could Connect 200k Unsolved Murders
< >
Image credit: YouTube/Outer Places

Crime shows like Numbers, Person of Interest, and even Sherlock portray criminal investigations as a kind of math puzzle: all the evidence is there, you just need to figure out how to parse the data. But reality is much more complicated than that—real cases are messy, and data doesn't just come neatly packaged for investigators to sort through. There's no mathematical formula for understanding the mind of a killer, right?



Well, yes. Now there is.



Thomas Hargrove was a newspaper reporter for over 20 years, and his primary beat was murder. One thing he and his colleagues struggled with was why so many homicides went unsolved; according to Hargrove, there have been over 200,000 unsolved murders in the United States since 1980, and many are never reported to the FBI or Justice Department.

 

Hargrove's interest in unsolved murders led him to create the Murder Accountability Project, now the largest database of murders in the United States. But Hargrove didn't stop there—he used all of that raw data to start building an algorithm that can identify previously unknown serial killers based on just a few points of data.





Years ago, Hargrove came across the phenomenon of "linkage blindness," a situation where various police departments or agencies have the data they need to catch a criminal, but because the information isn't shared across all the parties involved, no one has the complete picture. Two investigators might not even realize that their murder cases are linked unless they happen to spot the commonalities, Hargrove says.



To solve this problem, Hargrove created an algorithm that takes key pieces of information from a murder case, including the location of the murder, the victim's age, sex, and race, and the method used to kill them, then uses all those variables to create a nine-digit number. This number is then used in a technique called cluster analysis, which shows where similar murders have occurred. This led Hargrove to find a cluster of murders in Gary, Indiana, which he suspected was the work of a then-unknown serial killer. Hargrove approached the local police department with the data, but was met with silence. Seven more victims were found before the serial killer was captured—Darren Vann.



Hargrove's new method may revolutionize the way detectives solve cold cases and approach murders—instead of doing things by the book, they can start solving cases (puts on sunglasses) by the numbers.

Science
Science News
Technology