Putting the bite on crime with predictive policing

Imagine thwarting a crime even before it occurs by predicting when and where it will happen. Shades of the 2002 Tom Cruise futuristic action thriller Minority Report? Well, in a way yes. No ‘precogs” are being tethered to a computer yet, but some police agencies south of the border are experimenting with analytics technologies and what is being called predictive policing.

For example, for the last seven years the PredPol algorithm developed by a team of mathematicians and social scientists at the University of Los Angeles (UCLA), University of Santa Clara and UC Irvine, has been analyzing crime incident information to some members of the Los Angeles Police predict where certain types of property-related crimes are most likely to happen during certain police patrol shifts.

Similar experiments in police precincts in New York and Boston using other technologies have also been carried out.

As the PredPol Web site states, its mission is simple: “Place officers at the right time and location to them the best chance of preventing crime.”

The technology is designed to predict where crimes are likely to occur in place-based “prediction boxes” as small as 500’ x 500’.

PredPol is a secure, cloud-based, software-as-a-service (SaaS) platform. Crime data, specific to crime type and specific to police shifts is mapped into it instantaneously as the information is reported to the system.

Its developers say PredPol is built around the same technology used to predict earthquakes. Crimes, they said, occur along fault line and happen in predictable patterns much like earthquakes.

The PredPol report is shared is distributed to police during the roll call and call also be delivered to officers in the form of a paper report on transmitted to their smart phones or tablet devices.

While traditional crime analysis tools such as ComStat provide a “rear view mirror policing” by mapping past crime to come up with “some hints at the future,” PredPol tells the police where crime is likely to happen in the near future. For example, the model can actually predict the risk of a crime in one location for the next 10 to 12 hours even if the latest crime reported occurred in a different location, according to users.

In a six-month, random trials the system predicted twice as much crime as an experienced crime analysts.

In the four months following the technology’s roll out in the Los Angeles Foothill Division, crime went down by 13 per cent in the precinct compared to a 0.4 per cent increase in the rest of the city where PredPol was not used.

“This is the next era of policing…very soon we will be using a predictive policing where, by real-time crime patters, we can anticipate where a crime is likely to occur,” said William Barton, police commissioner of the LAPD.

And that’s among the chief worries of civil liberties advocates.

Predictive policing brings up concerns of class and racial profiling, according to Andrew Ferguson, associate professor of law at the University of the District of Columbia.

“Profiling suspicious act activities in not new, and it mirrors the daily practice of police officers who informally have an idea of stereotypical criminal activity,” he wrote in his paper titled Predictive Policing and Reasonable Suspicion. “…What is significant is that predictive policing will change the Fourth Amendment calculus of reasonable suspicion.”

For example he said, if a police officer sees a man loitering in a corner with a duffle bag and looking at a house, a police stop on that person based on reasonable suspicion would be difficult to justify in court. There is nothing objectively criminal about waiting with a bag, he said.

In another instance, if the police officer were informed via predictive policing algorithm that statistically a burglary were likely to occur in that area and police were told to be in the lookout, “a stop based on reasonable suspicion would likely be upheld.”

Ferguson worries that as predictive policing becomes more prevalent and begin to cover other crimes “predictive policing tips” may require less corroboration than other tips. This, he said, poses a looming problem for judges and lawyers.

He said the use of predictive policing in Los Angeles and Seattle may bring up concerns of racial profiling but the connection is indirect as the systems profile area and not the individual.

Predictive policing that focuses on people and behaviour rather than geography is more “troubling.”






Would you recommend this article?


Thanks for taking the time to let us know what you think of this article!
We'd love to hear your opinion about this or any other story you read in our publication.

Jim Love, Chief Content Officer, IT World Canada
Nestor E. Arellano
Nestor E. Arellano
Toronto-based journalist specializing in technology and business news. Blogs and tweets on the latest tech trends and gadgets.

Featured Download

IT World Canada in your inbox

Our experienced team of journalists and bloggers bring you engaging in-depth interviews, videos and content targeted to IT professionals and line-of-business executives.

Latest Blogs

Senior Contributor Spotlight