Predictive Policing

MANCHESTER – Police and policy makers – including the white house drug czar – were in New Hampshire Monday trying to find ways to tackle the heroin crisis.

To really understand it, you need to see it first-hand.

That’s where the Manchester Police Department comes in.

They offered a walking tour of the city’s heroin hotspots. Including a downtown area where three active drugs houses were recently shut down.

And now, the police has a new weapon in its arsenal to fight crime.

30-year-old Officer Matthew Barter fights crime in the Queen City, but his weapon is a computer, not a gun.

He’s a Manchester Police Department crime analyst.

His strategy – stop the crimes before they happen using what’s called predictive analytics.

In the 10 weeks the program’s been running, the police said robberies are down 24 percent. Burglaries are down 13 percent, and motor vehicle thefts are down 34 percent.

“It’s the same technology that Amazon uses to tell you what you should purchase next CUT,” Barter said. “That’s what we’re using to leverage in the policing world to identify where crimes are going to happen.”

The right areas may seem obvious – but the timing? That’s the science of stats.

“We take the city of Manchester, and we grid it out into five hundred square foot blocks,” Barter said. “For each one of those blocks, we look at which of those crimes happened here in the past few weeks, what are all the crimes that happened here in the past few weeks.

Is there an uptick in overall crime in this 500 square foot block? What happened during the same time last year?

The results?

“This past week, we had 60 percent accuracy in forecasting for those crimes within 500 feet of where we predicted it.”

Chief Nick Willard says the proof is in the numbers.

“It’s cutting edge policing is what it is,” he said.

But for Barter, it’s all in a day’s work.

So today, when the talk is all about heroin, this is an important process for the police. About 80 to 85 percent of property crimes in Manchester are heroin related.

Comments are closed.

Up ↑