« Gary Has a Bad Day of Tasting
New Crate & Barrel Update for Quinn »
[ # ] Minneapolis Shotspotter Reporting Disparities
March 25th, 2008 under Crime, Minneapolis


The Minneapolis Shotspotter system is a bit over a year old, and seems to be a great success. However, the current shots fired reporting published by the MPD presents an inaccurate picture of gunfire in Minneapolis.

For those you of not familiar with the system, the City of Minneapolis purchased a crime fighting technology called Shotspotter from a company by the same name early in 2007. Shotspotter consistes of a series of microphones installed on lamp posts that detect gun shots and calculate where the shot was fired. This information is immediately related to precincts for immediate dispatch of officers to the exact location of the discharged firearm. Impressive stuff.

And it works. For example, the MPD has case studies (PDF) of arrests made following shootings that were never reported. Shotspotter heard the shot and police arrived on the scene quickly enough to notice a fleeing car, stop it, and make an arrest.

Here are the stats from the first month of 2007 according to the StarTribune:

ShotSpotter has triggered dispatches to 69 suspected gunshot locations. The results: six felony and misdemeanor arrests, two recovered guns and a recovered stolen car. Information from ShotSpotter also is being used in homicide, robbery and shooting investigations.

Shotspotter is currently installed in 4 square miles of Minneapolis where the most gunfire has happened in recent years. As you can probably guess, this includes areas of South and North Minneapolis.

The Minneapolis Police Department publishes weekly Shots Fired Maps within the statistics section of their website. The maps include shots called in by residents, shots detected by Shotspotter microphones, and locations where bullets ended up hitting a person (aka. a shooting).

I’ve put together a slideshow of the weekly shots fired reports for the first 11 weeks of 2008 below. Red dots mark shots fired while red dots with a black dot in the middle signify shots that were detected by the shotspotter system:


Created with Admarket’s flickrSLiDR.

Unequal Reporting

Since ShotSpotter is only deployed in 4 of 54.9 square miles of the city, the shots reported on the shots fired maps are disproportionately in those 4 square miles. This makes those areas of town appear much more dangerous relative to other Minneapolis neighborhoods. This isn’t to say that the data is being over reported in Shotspotter areas. It’s just underreported everywhere else.

A Shotspotter press release from 2006 explains the increase in detected gunshots with Shotspotter:

“Glendale, Ariz, police began using ShotSpotter last year and detected 80 percent more incidents of gunfire than were reported by the public.”

If the same level of detection holds true in Minneapolis, the areas of South and North Minneapolis where Shotspotter is installed are getting hit with 80% more dots on the maps than they otherwise would.

The StarTribune’s article from Jan 2007 mentions that the MPD gets 5000 “shots fired” reports a year with half of those coming from the areas were Shotspotter is now installed. If that’s the case, the number of reported shots fired has likely jumped by 2000 in the Shotspotter areas of Minneapolis while remaining the same in the rest of the city. Does this also suggest that 2000 gunshots go unreported in Minneapolis every year in areas not covered by Shotspotter? I get that impression.

It seems to me that the areas of South and North Minneapolis have enough challenges without the police department publishing maps that give a statistically incorrect view of the city’s gun shooting problem.

How could this be solved?

1. a. Don’t intermix Shotspotter and called in shots on the same map.
1. b. Further differentiate Shotspotter shots on the map.
2. Outline the areas covered by Shotspotter.
3. Deploy Shotspotter city-wide to provide consistent reporting and crime fighting capability. (I’d like police to immediately know when a gun is fired in my neighborhood too.) Based on the reported cost of the first 4 square miles deployed, covering the entire city may cost up to $4 million. I’m sure someone from Shotspotter could give us an estimate.


Related Posts

  • Minneapolis Crime Watch on North Minneapolis ShotSpotter
  • ShotSpotter Spots a Shot
  • Minneapolis Crime of the Week
  • Political Reporting: SportsCenter Style
  • Baghdad Green Zone
    		•

    Read the Comments

    [ # 4959 ] Comment from justin [March 25, 2008, 11:22 am]

    The areas the shot spotters are installed are known to be the areas where more gun violence occurs. This is borne out by looking at the blue stars on the map. They occur more frequently in the areas where there are also shotspotter spottings. People also don’t always call the cops when they hear a gunshot. Shotspotters don’t have that limitation.

    While I’m sure it would be nice to have shotspotters everywhere, I’m pretty sure we don’t need them in Linden Hills, for instance. I think the cost to benefit ratio on blanketing the city with them is probably not worth it.

    Putting a little yellow line around the areas where the shotspotters are installed would be nice, I agree. But I can see why the cops would not want that to be publicly known information, despite the fact that it should be.

    [ # 4960 ] Comment from kohler [March 25, 2008, 1:03 pm]

    Justin, I absolutely agree that the Shotspotters, given a limited budget, are well placed to give the biggest bang for the buck. However, it’s unfair to those communities to report twice as many shots on Shots Fired maps due to better collection since it gives the perception that those communities are twice as dangerous as they actually are.

    It’s a side effect of otherwise great technology.

    Linden Hills has a reported shot fired between the 22-28th of January. I don’t know the story behind that report, but I imagine Linden Hills residents would like fast response times at times like that from police who wouldn’t normally be in the neighborhood.

    [ # 4969 ] Comment from ddjohnson [March 26, 2008, 9:17 am]

    A non-pdf version would be nice. imo.

    Write a comment