IPVMU Certified | 07/30/13 02:04pm
I have contacted ShotSpotter on numerous occasions, but they have always declined more than superficial contact.
I know the city of Baton Rouge, LA uses ShotSpotter and is satisfied with performance after some trouble getting the mics installed in the correct locations. There was an amount of moving equipment around because of echo effects and false alarms. They did not integrate the system with PTZs.
IPVMU Certified | 07/30/13 02:20pm
It always seemed kind of silly to have a system that detects gunfire. At best, it tells you a gun went off AFTER it went off.
"It always seemed kind of silly to have a system that detects gunfire. At best, it tells you a gun went off AFTER it went off."
True, but it takes the average gang-ganger about 15 shots to hit anything near their target, so the police have adequate time to respond.
Type in "shotspotter fail" into google search (or click the link, I've done it for you)
Personally, I don't see the value of this product at all....
As Mr. Rhodes points out, it is an after-the-fact tool.... the only first responders who could make it to the location in time to nab the shooter would probably be close enough to hear the shot themselves. Unless the shooter sticks around (as Mr. Karas sardonically points out) what is the value of the tool?
Beyond this innate flaw, what about false alerts? (read the stories)
Unlike on TV, all gun reports do not sound the same. Add in all the other similar-sounding reports (car backfire, firecrackers, indiana jones cracking his whip, guy clapping two bricks together, etc) and you've got cops running all over the place responding to shotspotter alerts.
Maybe if you combine Shotspotter with AISight, you could predict who is about to shoot a gun... :)
IPVMU Certified | 07/31/13 12:21am
The last thing law enforcement eagerly runs in the middle of is a gunfight. Presuming you have a raging gun battle in the streets, with endless bad misses and enough bullets to keep the bad guys going for enough time that an officer responds: what then?
I know exactly 0% of LEO training is devoted to breaking up a gunfight. They might try to disarm one before it starts, but there isn't a cop alive who is just gonna jump in to a bad crowd and start shooting in the hopes it breaks up an already violent situation.
No, they'll wait until the shooting stops, and then they move in. Either to chalk up the outlines, or start filling out the reports.
Shotspotter potentially increases efficiency in finding the exact location shots WERE fired, but does nothing to deter or diminish crime.
I've listened to a ShotSpotter pitch before and their focus was mainly on dispatching police to the location of a shooting to look for victims or evidence if I'm not mistaken.
"car backfire, firecrackers, indiana jones cracking his whip, guy clapping two bricks together, etc," according to them Marty, it can distinguish between all of those and provide alerts accordingly.
They use ShotSpotter in DCa nd I'm not saying it works, but I am saying, cops show up in force for gunshots, but not for firecrackers. That could just be that most the people in the neighborhood know the difference and don't call for firecrackers.
Interesting point, Carlton - citizens generally (not always) call the cops when they hear gunfire close by.....
I'm going to invent software that automatically plots the locations of 911 calls reporting shots fired that come in all within a short time period. It will plot all the geographic locations of the 911 callers on a map, and then dispatch cops to the center of all the plotted callers locations.
...anyone know any angel VC investors?
IPVMU Certified | 07/31/13 08:56pm
Is there any other option to these company? I think is really a good thing to add to the camera sites in city surveillance. And really easy to deploy over the existent infraestructure (you already have data connection and electricity).
Hernan, I have not heard of a direct competitor to ShotSpotter. Does anyone know of any?
Anyone know what the associated costs are for deploying this product?
Hernan: "I think (it) is really a good thing to add to the camera sites in city surveillance."
Can you expand on that statement? Why is it a good thing to add for city surveillance?
$100k wouldn't be especially bad per square mile. For instance, Manhattan is 22 square miles, meaning $2.2 million to cover the whole island. However, I have to imagine building heights and obstructions make deployment significantly more expensive / complicated. Can anyone comment on that aspect?
Yeah, there are going to be different mounting height requirements (mounting on top of a 33 story building isnt going to work). The deployment we did was a point to multipoint line of site wireless transmission backbone (Motorola) to Church Steeples, 3-4 story Warehouse rooftops, etc. We had 13 sensor locations covering a roughly 5 square mile area, but this was a typical old NY trade city, lower level buildings, a lot of warehousing, the tallest buildings in the city were churches and a single 19-story apartment building. It was a pretty ideal geography for the deployment, requiring a relatively low concentration of sensors.
Here is an efficacy study from 2011 on ShotSpotter found that officers trust ShotSpotter over 911 reports of gunshots. I think that says something. It's definitely worth a read.
It's no secret that ShotSpotter sends back a lot of false positives, but I guess a good question is whether it sends back more false positives than you're getting from the general public. How many people know the difference between a firecracker and a gunshot? An image from that study:
The general tone of that study was that it may have false positives, but it's better than the inaccurate information they get from callers.
Some other things, ShotSpotter acquired its main competitor back in 2009, and this story from the Post in 2007 shows how it helped piece together a shooting in DC. Also, here's a piece on the origins of this type of technology. Keep in mind similar technology (triangulating location using sound) has been around for YEARS.
In 1998, the idea was to use it for counter-sniper operations and in 2003 came the Boomerang.
I guess I have a problem with that study... Just as it has false positives I assuming it misses instances as well.. What if a truck is near the microphone, or a chopper overhead... What you are painting is a "we received several calls of gun shots but our computer told us not to respond". I don't mean to be all Will Smith here, but I'm not trusting those machines If your going to respond to every call anyway, and I would think you have too, right? What's the point of the system?
Chris, I'll play devil advocates here. Let's say, in a city, ShotSpotter falsely identifies 10 shots each month and misses 2 shots (real shootings are likely less than things that sound like shootings). But let's say by comparison, people calling in falsely identify 50 shots each month while missing 10 shots. You might still argue that ShotSpotter is a net positive because it gets you more real shots and lets you wade out more false calls.
I think that's the point of the system - that it would be better than the status quo - which is random call ins.
I don't have any experience with ShotSpotter so I don't know how well it works but that's the type of argument one would make for an imperfect technology (that it's better than the existing situation and therefore delivers value).
Sorry John, I have to disagree... Your example is putting the potentional for a loss of life at the same value of the cost of an officer responding to 5 false alarms... If someone is dead, or a family is dead because the police didn't respond to the only caller, there is in no way a positive
Also, I agree with anyone who says not picking up sounds of gunshots is a problem.
But I do wonder how many of those false positives are user error. This city used ShotSpotter to make recordings and then a dispatcher would decide of the recording was a gunshot or not. They had a 95 percent false positive rate. Additionally, false positives are supposed to be classified in the system as such so the system can "learn" and get more accurate. From this article about the report mentioned above:
"The report also noted an almost universal lack of reclassification of false positives. After determining that a ShotSpotter activation was not a gunshot, users are supposed to reclassify it as a false positive in the system. This both makes the statistics more relevant for analysts and helps fine-tune the learning algorithm that determines what constitutes a gunshot. By reclassifying the activation and improving the algorithm, users can reduce false positives without increasing false negatives and make the system more valuable overall."
IPVMU Certified | 08/02/13 05:22am
If this technology works, and you have a city surveillance solution already deployed. Then, is not hard (maybe expensive) to add this feature to your installation.
It looks like we don't have any options then...no other "sound triangulation" technology around.
Do you really think is possible to ask the operator to "teach" the system to recognize the real shoots?
Trafficware, a CUBIC Company
I've worked with it in the past, and it was pretty good. A few false positives, but good overall. There are also software companies out there that build enhancement systems to add feature and fuctionality.
The last I played with them was maybe 3 years ago, when I was working for an integrator focused on city-wide mesh based surveillance. SS was in the middle of changing their model from mesh based network to 3g/4g, which ticked us off. Before that, we were involved in deploying SS for several major cities as well as for several large public events, and we were pleased with the results overall.
Hal, nice feedback, thanks! Why were you 'ticked off' about them changing from mesh to 3g/4g?
"If this technology works, and you have a city surveillance solution already deployed. Then, is not hard (maybe expensive) to add this feature to your installation."
I'm not saying it is difficult to deploy, nor do I really care about false positives (though, how are false positives even ruled false? answer: by currently available means, i.e. without the aid of this 'technology'). wrap your head around that, :)
I'm trying to understand the 'value' of paying that much dough for a 'technology' that helps LE do what it appears to me they can already do just fine.
Hence my question, why is it a good thing to add to a surveillance system? Are municipalities finding bodies all over the place where the general public has not alerted them to shots being fired?
I always assumed that mob guys in Vegas took people out of the city to the desert to whack them so nobody would hear the gun shots...
I have never been impressed with these systems. I assume it depends upon what the municipality is trying to accomplish, but, would it not be a better investment to cover these areas with PTZ cameras (which have audio features depending on legality of use in the jurisdiction)? I think for what you would spend on the Shot Spotter and the long term maintenance and ongoing support, you could have a more effective crime fighting tool with a network of IP cameras.
I'm always looking for unique sensors/effectors that we can use. When I contacted ShotSpotter a few years ago, I found the sales rep assigned to my area to be arrogant. I don't know if that's because we were a small company.
I found a company in Tucson, AZ called Safety Dynamics that is more willing to work with small companies. Unfortunately, our focus and priority has shifted and we haven't had a chance to revisit with them.
I don't personally know how well these things work; I happen to work with a partner who has contacts in Chicago and I understand they like what they have. I guess that means it works for them. We have not had the same interest in our West Coast first responder contacts.
Somehow I fail to see how it can significantly reduce the number of shots fired. 94% reduction is a huge number considering these sensors are not, by their nature, designed to deter much less prevent crime. Does this mean they experienced significant reduction in crime committed by guns as well, and if so, by how much?
Check out this Nassau County legislature guy, seemingly falling-down drunk on Shotspotter koolaid before it was rolled out.... best quote:
"I am encouraged by its effectiveness from data in other cities where it’s already used showing that gunfire rates have been reduced by nearly 60 to 80 percent and violent crime by up to 40 percent."
Followed by the money line: "It is my deepest hope that just knowing that ShotSpotter is out there will be a deterrent to using guns, and that it will also help remove illegal guns from the streets, homes, and lives of everyone in our neighborhoods and make a safer environment for our children and their families.”
Elegant politician-speak, no doubt. But hope don't pay the bills, Kevan.
I find the last line in the Suffolk County link above especially enlightening:
"Continued monitoring of the . . . system through the end of 2013 would give ShotSpotter time to begin . . . improvements and provide us with enough reliable data to perform an analysis of system benefits in conjunction with current costs," the report concluded."
um. Aren't cost/benefit analysis studies usually what folks perform BEFORE they spend $800K?
I guess they were 'hoping' like Kevan was back in 2009.
Criminal: "Oh shit, they have SpotShotter, I am not going to fire guns anymore." This seems unlikely.
Also, the problem with those stats is controlling them for other factors, i.e., violent crimes have fallen for years in the US with or without SpotShotter.
Netlogix is their closest competitor (although neither will probably admit it). This technology has been in use for some time in the military. Not sure which one (if either) supplies it. I also think you may see it rolled out for schools in the near future.
Although the WP article says no empirical research has been done, the U.S. DOJ commissioned a Field Evaluation study in Redwood City (although it is dated) which you can review here.
Having been in policing for 26 years, I would ask myself, how many police officers or surveillance cameras might a city have paid for with the 5.5 million that Washington D.C. has invested in this system since inception.
Another article about SpotShotter in Newark . Not the best review, but informative. It sounds like more of a "deterrent" and not all that effective to John's point ""Oh shit, they have SpotShotter, I am not going to fire guns anymore." I'm not sure if it's the implemenations, the maturity of these types of products, but there seems to be some very mixed views out there. It certainly seems very expensive for the results provided so far.
In response to Jim, I work for Thales and I'm pretty sure we deployed Gunshot Detectors in Mexico City. I'm going to see how effective they are and any feedback (and try to not be biased). I'm not sure how much of a product line it is for Thales. Admittedly, I only learned about it because of my involvment in some of the other security aspects of the project. I'll see what I can dig up.
Gunfire-Detecting Technology Improving Public Safety in Pittsburgh’s Most Dangerous Areas
City Council hopes to extend the contract with SST to continue using the ShotSpotter technology.
Pittsburgh started testing the system in 2014 and now the city wants to extend the contract with SST through 2016 to continue to improve public safety. So far, the city has paid the company $194,000 for the installation of the equipment, and it would cost an additional $135,000 to extend the deal through the end of the year.
IPVMU Certified | 04/09/17 01:57pm
Still the same? Is there more information about this?
For those of you that don't know me. I recently retired as a Police Detective from a 380-500 Officer Department in South Jersey. I also have a NJ state licensed security installation company for the past 23 years. From my experience we had shot spotter and it worked very well. They are in the process of intergrading it into a new realtime crime center where they have Officers watching 100's of live feeds. The system is amazing how fast and accurate the information is. We rearley ever get a false report. But there were times where we did not receive any shot spotter alerts and we had victims shot or shell casings recovered. If a small caliber handgun was used inside a dweling or vehicle and the windows are closed and a sensor is not in range it will not register and trigger off the shot spotter. On many of cases by looking at the shot spotter in the field we were able to locate a crime scene which produces the evidence. Shell casings, Gun, victim, finger prints etc. The system is so sophisticated if multiple guns are shot it will mark the area where the shooters were at the time of each shot. Which if you have video in the area you already have time,location so you can check for video evidence. There has also been about eleven arrest of suspects with weapons because of how quick ShotSpotter is.