Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ShotSpotter: listening in on the neighborhood

computer.rip

Post reply on HN
Print view View original

411–420 of 443 posts

Re: ShotSpotter: listening in on the neighborhood

#411

Post by bastawhiz »

assimpleaspossi wrote: post #84
bastawhiz wrote: post #35

Earlier quoted context omitted.

> It just says a signal if there is a loud noise in public. This is probably false. There have been two convictions as a result of recordings of conversations that happened through the devices. I, for one, do not want microphones capable of recording my conversations placed near me by the government. Not because I'm paranoid the government is conspiring against me, but rather because I have zero confidence they're st…

It sounds like you are paranoid. Why would the government be listening to a conversation you might be having near a device used for tracking gunshots? Is there some little man somewhere just waiting to pounce on you or anyone just for that purpose? I don't know how effective these devices are but any tool that helps the police go after criminals is a good thing and no one should be fighting against good things.

You clearly didn't read my comment. We know the recordings exist, that's indisputable. How securely are they stored? How do I know the government isn't selling those recordings to third parties to make a quick buck? How would you know that a conversation you had with nobody around isn't going to leak online and be scrutinized by god knows who? I don't care one way or the other what the government thinks of what I say, but I don't trust them to secure their systems and protect my privacy more broadly, and you shouldn't either. The system could just as easily do exactly what it says (detect gunshots) and NOT save recordings, but that's not the case.

Re: ShotSpotter: listening in on the neighborhood

#412

Post by bastawhiz »

assimpleaspossi wrote: post #374
humanistbot wrote: post #162

Earlier quoted context omitted.

Over a two year period in the US, there were over 600 cases where police officers and civilians with access to law enforcement databases violated internal policies and safeguards to access private data about "romantic partners, business associates, neighbors, journalists and others for reasons that have nothing to do with daily police work". This is certainly an undercount, as this is just who got caught and the depa…

The same thing may have occurred with an office just sitting in his squad car on a street corner. Will we now pull them off the streets for that reason? Will we require them to keep their windows rolled up or wear noise cancelling ear muffs? The cases you cite are of people who did things wrong. The cases you cite have nothing to do with the advantages this gives to the police in protecting citizens and solving crime…

> Lest anyone forgets, the police are on our side.

This wouldn't be quite so laughable if civil forfeiture didn't exist.

Re: ShotSpotter: listening in on the neighborhood

#413

Post by bastawhiz »

curtisblaine wrote: post #42
bastawhiz wrote: post #29

Earlier quoted context omitted.

> more gun crime equaling more surveillance But if you're not listening for the guns then you're not going to find the crime that does happen, will you? It's either saying "we don't care about it because it's less frequent" (which is stupid—you still build a fire station in places where buildings burn down less frequently) or "it's easier to keep ourselves busy when we fish in a barrel".

If you were to design a system to spot a certain kind of crime in real time, would you place it in zones historically more plagued by that kind of crime or in zones that experienced it scarcely? It's hard to buy into the rhetoric of "it's only fair if you monitor the crime equally" if you have already independently sampled it for decades and by now know the zones where it's much likely to happen. Would you patrol a d…

> would you place it in zones historically more plagued by that kind of crime or in zones that experienced it scarcely?

Both. I don't skip out on adding monitoring to my applications that don't crash as often. The crashes might be less frequent but that doesn't mean I don't need monitoring.

Re: ShotSpotter: listening in on the neighborhood

#414

Post by bastawhiz »

dkarras wrote: post #86
bastawhiz wrote: post #29

Earlier quoted context omitted.

> more gun crime equaling more surveillance But if you're not listening for the guns then you're not going to find the crime that does happen, will you? It's either saying "we don't care about it because it's less frequent" (which is stupid—you still build a fire station in places where buildings burn down less frequently) or "it's easier to keep ourselves busy when we fish in a barrel".

you can definitely tell gun crime has happened without listening to the sounds of it. people die, get hurt, police reports are made etc. we don't typically determine if gun crime has happened by... listening do we?

Then why do we need it anywhere?

Re: ShotSpotter: listening in on the neighborhood

#415

Post by hcfman »

nonrandomstring wrote: post #371
hcfman wrote: post #268

Earlier quoted context omitted.

Categorisation will certainly be error prone as there is so much similarity because fireworks and gunshots. The main difference I expect is the speed of the shockwave resulting in steeper spectrograms for gunshots. However, a steep spectrogram is essentially showing a large amount of high as well as low frequencies. The higher frequencies degrade very quickly over distance reducing the difference of a gunshot from a…

Nice work. PM me if you want to chat. Have you considered the possibilities with oversampling and adding more channels- now there are much cheaper 192kHz multi-chan ADCs on the market - these can be used with MEMS arrays. Then you can play with phase information much more freely.

Will do. Interesting looking show you run btw.

Re: ShotSpotter: listening in on the neighborhood

#416

Post by zemo »

fasthands9 wrote: post #194
zemo wrote: post #81

Earlier quoted context omitted.

> It just says a signal if there is a loud noise in public. as a community we really ought to be suspicious of claims like this. There's another word for "acoustic sensor" and it's "microphone". When a corporation says "trust us, we only use the data for good", we should collectively raise our eyebrows, knowing that when data is collected for one stated purpose and stored, that the data at rest will always be used fo…

Being skeptical of face value claims makes sense if they have never been used before. But, they've been in place in some cities since before 2000 and after searching I'm still not aware of any incident audio from it was used that didn't have to do with a gunshot. Looking into it more all the cases of audio being used in court, that was not audio of gunshots, was audio immediately after a gunshot. Spotshotter is open…

> I know we likely disagree on the usefulness of those, but if you are privacy concerned why would shotspotter even be on the same level of concern as cameras that don't need a gunshot to start filming?

I think that is also bad. Two things can be bad! That's just not the topic of the post.

Re: ShotSpotter: listening in on the neighborhood

#417

Post by ALittleLight »

hcfman wrote: post #410
ALittleLight wrote: post #114

I've also been blogging about ShotSpotter and review the evidence that the system is good or bad. https://quickthoughts.substack.com/p/shotspotter-good-or-bad There's a lot of evidence that ShotSpotter detects almost all gunshots and that's been validated by multiple third party groups. ShotSpotter also seems to alert to things that are either false positives (construction sounds, fireworks, etc) or are not useful. C…

Fine but it’s about time there was external validation possible from the public and transparency about the math.

In some cities there is - kind of. DC, for example, publishes all ShotSpotter events publicly and you can download all records of their events. You can notice by plotting their events that there is a lot of extra "shots fired" on January first (probably from fireworks). If you knew of gunshots at a specific geographic coordinate you could confirm or deny with the DC data. I suppose you could also go to some specific spot in DC and fire a gun to see if they could detect it - but I don't recommend that.

https://opendata.dc.gov/datasets/DCGIS::shot-spotter-gun-sho...

Re: ShotSpotter: listening in on the neighborhood

#418

Post by jncfhnb »

ta988 wrote: post #398
jncfhnb wrote: post #363

Earlier quoted context omitted.

Police presence increased the detection of crime and thus increases the statistics of crime in the area

So the bodies of people shot are seen more often in high police presence areas and just not found in other areas? That seem strange that it would be due to higher police presence. The fact that for non violent crimes the stats are skewed sound plausible but that seem less so for violent ones.

Dead bodies appear in about 1% of violent crime.

Re: ShotSpotter: listening in on the neighborhood

#419

Post by BriggyDwiggs42 »

whatshisface wrote: post #260
BriggyDwiggs42 wrote: post #155

Earlier quoted context omitted.

Actually this is a direct example of how systemic racism works. The goal isn’t to surveil minorities more, it’s just that the system watches places where crimes occur. Consider that, when an area has more surveillance, more people who commit crimes are caught. All of a sudden, you’re disproportionately catching minority criminals. Better put in more surveillance to cover the higher crime rates in those minority commu…

> Consider that, when an area has more surveillance, more people who commit crimes are caught. All of a sudden, you’re disproportionately catching minority criminals. Turn the coin over to the other side - you're catching criminals victimizing minorities. Nothing is simple, is it.

Uh sure, but that doesn’t effect the placement of new surveillance systems, which is the reason I mentioned it. Also, I think you’re sort of saying that because you think catching them is a huge success, and that that offsets the harm, but it’s not gonna do a ton of good when nothing has been fixed about the underlying causes. At best the victims get catharsis, where if that money imprisoning the criminal were spent on something like providing economic opportunities or funding schools they’d likely benefit more and be more secure against future crime (not that its either or).

Re: ShotSpotter: listening in on the neighborhood

#420

Post by jjjjj55555 »

HDThoreaun wrote: post #94
larrymyers wrote: post #28

Earlier quoted context omitted.

Chicago is letting the ShotSpotter contract expire instead renewing this year after the data revealed that it did not have any positive effect on crime reduction. https://www.yahoo.com/news/chicago-not-renew-shotspotter-con...

I think that's because our police force is incompetent though. Shotspotter may very well suck but if the police cant solve shit either way its not shotspotters fault. Murder has a 5% clearance rate here

Are they incompetent or just aren't even trying? I ask myself this question all the time.
Post reply on HN