The NYPD has a new algorithm to help it fight crime
The AIs are coming for you.
The New York Police Department has a new tool in its investigation of crime, a pattern-matching software called Patternizr, which reduces the search for crime patterns to a click of a button.
The Associated Press (AP) reported on the new tool on Sunday, March 10, which took two years to develop in-house before it was rolled out to all the NYPD’s precincts in 2016. Before you start worrying about Minority Report-style pre-crime arrests, Patternizr isn’t built to forecast crimes. Instead, it taps into the hundreds of thousands of prior crimes that the NYPD has logged in its database, surfacing patterns of behavior to tie in with new crimes.
Before this, analysts in any of the department’s 77 precincts mainly looked at prior crimes in their own precinct, making it harder to catch repeat offenders if they did their crimes across the city. That also enables the department to move more quickly against criminals, as Patternizr can find those patterns far more quickly than a manual search by analysts.
A decade of patterns
Patternizr was trained on an analysis of 10 years of crime patterns that the department had created manually. According to NYPD assistant commissioner of data analytics Evan Levine and the former director of analytics Alex Chohlas-Wood, who was behind the creation of the program, this is the first system of its type used by a police force in the U.S.
They told the AP that in testing, “it accurately re-created old crime patterns one-third of the time and returned parts of patterns 80 percent of the time.” Just like when a human crime analyst tries to identify crime patterns, Patternizr compares factors such as the method of entry, type of goods taken, and the distance between crimes. The creators also drew on the wealth of knowledge from the uniformed officers who had decades of experience identifying these patterns using the traditional methods.
The example used in the AP report was of a “syringe-wielding drill thief [who] tried sticking up a Home Depot near Yankee Stadium.” You’d think that with a crime that specific, it’d be easy for analysts to check if any similar incidents had occurred. That’s not necessarily the case according to Bronx crime analyst Rebecca Shutt, who worked on the Home Depot case, saying “because Patternizr picked up those key details in the algorithm, it brought back complaints from other precincts that I wouldn’t have known.”
While the system has been in service since 2016, the Patternizr software only became widely known thanks to Levine and Chohlas-Wood discussing their work in a recent article in the Informs Journal on Applied Analytics.
- Ring’s Neighborhoods Portal gives police access to loads of recorded footage
- DJI expands its geofencing to make sure people aren’t flying their drones around airports
- Self-driving cars can be commandeered by both hackers and police
- Mark Zuckerberg publicly posted his privacy thoughts
- Elizabeth Warren wants to break up ‘Big Tech’ companies like Google and Facebook