Is predictive policing right for your smart city?

Thu, 2016-11-10 14:54 -- Doug Peeples

Predictive policing, the process of using data and analytics to anticipate where and when crimes will occur, has taken a lot of heat from critics who say it encourages discrimination against certain segments of the population and racial profiling. We can't say that's never happened. But we can offer the reminder that it is a disruptive technology, much like smart meters were several years ago. And like most disruptive technologies that have arrived in recent years, predictive policing is evolving — getting better.

When utilities first began installing smart meters, they saw no need to explain why or what the benefits for their customers would be. An immediate backlash alerted most utilities that they should have taken steps to educate and engage with their customers first, and most took corrective action.

City governments and their police departments may want to take the utility experience to heart and explain to the citizens they're working to protect how predictive policing can provide them with a safer place to live. And most cities likely already have the means to engage with those citizens. We've written before about a platform from Council Associate Partner Sungard Public Sector that incorporates social media into police management systems. That integration gives citizens access to the information they want about police activities and, as a result of a more transparent and trusting relationship, citizens are more likely to provide tips that help police do their jobs. If your city is serious about becoming smarter and more livable, at least consider how predictive policing could enhance public safety. It's a smart move. And don't forget to keep your citizens in the loop.  — Doug Peeples

Data analytics has become an essential part of how smart cities operate, from managing traffic flow to making trash collection more efficient and cost-effective. In the past few years, it has become a key component of how law enforcement agencies do their jobs because it gives them a way to anticipate where and when crimes will are most likely to occur.

Many consider predictive policing to be the future of law enforcement. It's in use in police departments in several states and cities from Santa Cruz, California to London report significant reductions in a variety of types of crime.

As we said above, predictive policing technology does have its critics and it hasn't been a good fit for all law enforcement agencies that have used it. Earlier this year in California, Milpitas Police Department officials discontinued a contract with a predicting policing software company because they didn't see enough benefit to justify the cost. Burbank, California's police department suspended using it to deploy officers, but is still using the information it gets from its program. As Burbank Police Chief Scott LaChasse said in a Los Angeles Times news report, "The future of law enforcement is not going to be random patrol, it's going to be predictive analytics."

How does predictive policing analytics work?
Council Lead Partner IBM has been involved in the field for some time. Earlier this year, the company explained predictive policing this way. "This type of technology can gather data from a wide variety of sources, even beyond historical crime patterns. Digitized agency-provided information, like data from police blotters, emergency band communications, interagency feeds and camera feeds, is an essential input for predictive analytics. But as mobile technology and the Internet of Things connect more people an devices all over the world, they create very large quantities of data that can be factored into the predicting policing process. Much of this data can be gathered and processed in real time, leading to improved response times."

Council Lead Partner Microsoft also has been heavily involved. Its Azure Government end-to-end cloud solution was chosen by predictive policing solutions company PredPol, a company popular with law enforcement agencies. PredPol's system uses an algorithm to predict where and when crimes will probably occur based on criminal behavior patterns and provides maps to police to help them zero in on areas where there is a good likelihood that they can prevent crime.

Something else to consider
Predicting policing programs are generally not expensive. Part of that is because police departments usually have most of the tools needed already in place. And as IBM points out, the technology also enhances department efficiency in that there not as many fruitless officer and equipment deployments, fewer false alarms and cold cases.

For more on public safety...
Take a look at the Public Safety chapter of the Smart Cities Readiness Guide to learn about the importance of public safety intelligence (data) for police, fire and other emergency operations and how to generate, collect, evaluate and leverage it for better outcomes. The chapter includes recommendations and guidance as well as case studies that illustrate successful projects that have made cities safer and more livable.

Doug Peeples is a Portland, Oregon-based writer specializing in technology and energy. Follow @smartccouncil on Twitter.