Thanks for the note. It sounds like your spotshotter technology is helping the police arrive on the scene quicker, which increases the likelihood that a victims life can be saved. That is definitely a fantastic benefit of the system. From my surveillance research, I believe gunshot detection has the ability to have a bigger impact on a city due to the sheer size of the coverage area. In addition, costs are much smaller when compared that of a public CCTV system. In some cases, I have also read the private sector has funded your systems. All of those things are a win in my book.

As far as your comment on "not enough data", I am thinking the opposite. I see there has already been two academic papers written on your technology to date. According to the abstract, the first paper only used 2 months of data.

One could argue that these three cities comprise perfect datasets – one for the experimental location (East Orange), and two controls (Newark and Irvington). From what I understand, East Orange installed the system in 2003, Newark installed the system in 2007, and Irvington is currently underway.

The small number of murders, as you put it, spaced over a 10-12 year time series analysis would be dealt with by using the proper statistical methodology. Possibly even with the poisson method as found in the paper I co-authered (see news link on top).

There is a great paper written by Farrington and Painter which lays out how to perform a CCTV impact analysis. The same methodology and approach can be applied to any spotshotter installation.

As far as the link describing LA county and crime reduction, I am not disagreeing. However, until a thorough study has been performed which includes peer-review, claims of success can only be just that – a claim. A peer-reviewed paper could help your company point out the positive things like the lives that are being saved due to the rapid dispatch of an ambulance.

Great question. Unfortunately the up-to-date crime counts are not on-line. If you look at the crime counts on citi-data for East Orange, Newark, and Irvington, you'll see murders are higher for 2003 – 2007 in Newark and Irvington and slightly down for East Orange. I cannot find the crime count numbers on-line for 2008 and 2009 for all three cities. Also, those crime counts are murders, not specifically death-by-gunshots. They could have been stabbings.

So, until we can see the murder crime counts for 2003 to 2009, broken out by murder-by-gunshots, I believe overall murders for these three towns have not declined. I believe per capita, they may have actually increased.

Maybe even murder by strangulations have increased because criminals know they stand a good chance of getting caught if they use a gun.

I agree with your greater good theory, but unfortunately for the citizens of Irvington, they may have ended up on the wrong side of East Orange and Newark's gunshot social experiment. Mayor Cory Booker is a good man, but he is a politician on the rise. He will talk up the good aspects of this initiative (i.e. Murder in Newark is lower) and downplay any residual side effects that may be occurring in Irvington.

First of all, the "event-alert-intervene" flow is *PRECISELY* what gunshot location/detection technology most directly helps improve. Bear in mind that with the exception of one particular city, less than 50% of gunfire is reported to 9-1-1 in any city we're deployed in nationwide (we're in 45 cities). In many cities, that number is less than 20%. So the first step in improving the event-alert-intervene process is to alert on a higher percentage of gunfire incidents than would otherwise be alerted on. (That does increase the number of calls for service, but since the location is far more precise than a 9-1-1 call usually is, there's a substantially reduced workload for responding officers when they get to a reported incident.)

Second, the systems in New Jersey have, by in large, not been installed long enough to produce noticeable trends in murder rates, primarily because (thankfully) the murder rate is a "small n" problem (in other words, fewer than 100 incidents per year–thankfully!), so it takes a long time to measure improvement. If you're looking for data on reduced murder rates, I can point you to this article about a 40% reduction in homicides in Los Angeles (http://www.shotspotter.com/news/articles/2009/1-January/012309_LACounty.html) or this one about a 35% reduction in violent crime in South Carolina (http://www.shotspotter.com/customers/casestudies/cs_ncharl_630-0010-01-A.pdf).

Best,
-James

If so, even if one area slightly increase, if other areas significantly decreased, is that not overall better for the state of NJ?

I think displacementis an important issue. However, if crime goes from 100 and 50 in cities A and B to 60 and 60 after city A implements a solution, city B may be slightly worse off but overall the 'state' is better, no?