Arming Cops With Big Data To Predict—And Prevent — Crime
August 18, 2015
“You’re a disease,” actor Sylvester Stallone tells a knife-wielding criminal in “Cobra,” a cop film from 1986. “And I’m the cure.”
Sly probably didn’t realize it at the time, but his quote was prophetic—law enforcement agencies have begun viewing crime as a disease, employing the same data-driven tactics that scientists use to understand how illnesses move within populations.
Epidemiologists have long used data to create a statistical bird’s eye view that helps them predict where and how a disease may spread. Historically, this data was gathered in painstaking manual surveys, making it difficult to track fast-developing situations. Since then, epidemiologists have begun using digital data collections to improve accuracy and responsiveness. Websites like SickWeather, which uses mobile apps to map incidents of sickness, combine epidemiology and big data.
Big data strategies make sense for law enforcers too. As early as 1981, law enforcement agencies were working with epidemiologists at the Centers for Disease Control to better understand the causes behind the spread of violence.
Today, authorities around the world are analyzing crime data on a regular basis, using “predictive policing” to determine which areas may begin to see certain kinds of crimes and where more crimes might occur.
Data Becomes A Weapon In The War On Crime
When Stallone donned his cop shades in “Cobra,” the PC was only five years old, and data storage cost nearly $25 per megabyte. Today, that same megabyte costs just three-thousandths of a cent, making data storage and processing downright cheap. As a result, many law enforcement agencies are repeatedly turning to big data to learn new things about crime.
In New York City, the police department has progressed from using push pins on paper crime maps to putting digital markers on high-definition displays, helping them identify patterns more quickly and easily. The technology underpins the data-driven CompStat policing system developed by the department in the mid-90s and still used today.
By providing digitized crime statistics with detailed information about the crime location and type, CompStat has helped the NYPD deploy police officers to problem areas and hot spots with greater precision.
The Los Angeles Police Department is crunching years of crime reports in a system called PredPol, which spits out colored boxes on a digital map. The boxes represent predictions about where particular kinds of crime are likely to occur. These boxes move frequently based on new information, giving police officers a simple view of a highly complex set of indicators so they can better focus their efforts. A trial of PredPol has shown that the system accurately predicts twice as much crime as other methods.
Taking The Long View Predictive policing will continue to evolve as data, storage capacity and computing power become more plentiful.
At the Rutgers School of Criminal Justice, academics are attempting to identify the risk factors that cause crime to escalate and migrate. They’re plugging in the addresses of criminals, likely escape routes for perpetrators and where potential victims are likely to be targeted to develop new kinds of crime equations.
ShotSpotter, which uses sensors to detect gunfire in real time, not only alerts 911 operators, but also saves digital records for police data-crunchers to analyze later.
Social media “listening” tools, meanwhile, have provided investigators with valuable intelligence on gang structure and criminal intent, with the goal of better enabling authorities to stop crimes from happening.
In one example reported by Fast Company, Canadian cybersecurity and intelligence consulting firm SecDev helps U.S. government aid agencies understand what’s happening among drug cartels on the Mexican border, so they can better direct funding towards preventive activities, such as community policing and early-childhood intervention.
SecDev mines social media to learn about shifts in cartel power, mapping posts geographically and developing gang lingo glossaries to help law enforcement get a better understanding of what they’re up against. The company works with law enforcement agencies in San Diego to help map a constellation of gangs there and track their links to criminal networks in Central America.
These days, it is possible to treat crime like a disease, but just as healthcare and epidemiology have evolved, so has policing. When it comes to predicting crimes and saving lives and property, an ounce of prevention is truly worth more than a pound of cure.