What started as a peculiar crime quickly turned more sinister when a syringe-wielding man caught trying to steal power drills from a Home Depot in the Bronx attacked an employee with the hypodermic needle. The incident left the employee injured and could have left police stumped, but the NYPD quickly realized the suspect, who got away from the scene, had likely done this before. In another incident, a man caught shoplifting a drill had also waved around a hypodermic needle to threaten an employee at a Home Depot in Manhattan, police said.
The pattern was detected not by an officer combing through records, but by an analyst utilizing computer software developed by the NYPD to better fight crime, officials said. The software, known as Patternizr, allows the department’s crime analysts to easily search for similarities among burglaries, robberies and grand larcenies across the 77 precincts covering the five boroughs.
Historically, NYPD crime analysts had to wrack their brains and dig through reports to identify potential patterns of crimes occurring throughout the city. It was time-consuming, and especially difficult when a crime pattern expanded beyond the borders of a specific precinct, authorities said.
“It was a memory-based process … kind of a manual process and it was inefficient,” Evan Levine, the NYPD’s assistant commissioner of data analytics, told InsideEdition.com. “In addition to that … it was difficult for the department to identify patterns that crossed precinct boundaries.”
But Patternizr allows police to break down those precinct boundaries, Levine said.
The technology uses pattern-recognition algorithms to sort through thousands of NYPD records, comparing factors such as time of incident, method of entry, type of goods taken, if a weapon was used, and if so, what kind, and the distance between crimes.
The software allows analysts to spend “less time on the manual process and more time thinking about the actual crimes,” Levine said.
In the hypodermic needle incidents, the algorithm likely identified the matches “because of high similarity scores for the time of occurrence, the robbery sub-type (began as shoplifting), the weapon used, and several matching words in the narratives (e.g., drill, needle),” officials said.
Investigators combined the two complaints into an official pattern, and passed that information, along with two other larcenies committed by the same perpetrator, to the detective squad. A suspect was eventually arrested and later pleaded guilty to larceny and felony assault. He is currently awaiting sentencing.
Levine and Alex Chohlas-Wood, the department’s former director of analytics, developed the software for two years and it began being utilized in December 2016. The NYPD provided an overview of the technology in the INFORMS Journal on Applied Analytics last week in the event that other police departments want to utilize the software.
“We published all the details of the algorithm,” said Chohlas-Wood, who now serves as deputy director of the Stanford Computational Policy Lab at Stanford University in California.
The software does not track rapes or homicides, but Levine and Chohlas-Wood noted the department uses traditional resources to identify patterns concerning such crimes.
After Patternizr identifies a pattern, it undergoes multiple levels of vetting before being officially deemed as such, authorities said.
“There are many levels of supervisory review … to become an official, departmental pattern,” Levine said. “We need to check each other’s work and make sure we’re on the same page.”
Authorities also took into account while developing Patternizr the possibility of biases in the algorithm. Those imputing details of a crime are not able to include the suspect’s race or gender, and Patternizr does not examine for those details.
“There’s been a lot of concern in the public [about] the use of algorithms in criminal justice replicating bias if it’s there. We took a couple precautions to prevent this from happening with Patternizr,” Chohlas-Wood said.
Levine agreed, saying: “This is really about keeping New York City safe.”