Data-driven policing

This past week in Seattle the city government put on a press briefing to talk about a data-driven policing initiative, which aims, among other things, to deploy policing resources more effectively, including by predicting where higher level of crimes are likely to occur.

A white paper I found on the web describes the problem being addressed:

"With limited analytical resources, the Seattle Police Department relies heavily on its officers to determine where to patrol, based on their experience and intuition. This model of deployment has proven to be limited in its effect on reducing crime. Recent studies have shown that evidence-based patrol deployment improves the patrol officers’ ability to pro-actively affect crime."

It's interesting that Seattle's mayor, Mike McGinn, introduces a remedial facet to the initiative. He was quoted in a KUOW piece as follows:

“Our aim with data-driven policing is to minimize the influence that unconscious bias can have on these decisions [of where to deploy officers]. With this software we can take that out of the equation.”

If you don't live in Seattle, you may not be aware of the problem the City has been having with bias in policing and the contentiousness around a settlement between the City and the US Department of Justice. I'd infer that the Mayor's remark speaks to that context.

But I don't think that predictive software is doing the policing on the street.

James DermodyMost interesting to me on this topic were the remarks of Seattle Police Captain James Dermody, who spoke to the recent success of the Seattle Police Department and the community of the Central District in Seattle in reducing crime at the intersection of 23rd and Union. (This success was recently written up in a high profile story in the Seattle Times.) Capt. Dermody attributed the 80% drop in crime in the area to "data, smart tactics and leveraging our relationships with the community."

It sounds like the main function of the software is to help the Department make decisions about how to deploy officers during something called "uncommitted time," which represents 30% to 40% of an officer's available time. Once deployed, the officers still have to do the work of talking, investigating, building relationships.

You can see Capt. Dermody's remarks in this video from the Seattle Channel, beginning at about minute 6.


blog comments powered by Disqus