Slashdot Mirror
Twitter Detects Riots Faster Than Police, Study Says (cnet.com)
A new study by Cardiff University has determined that Twitter can be used to identify dangerous situations up to an hour faster than police reports. From a report: Researchers at Cardiff analyzed 1.6 million tweets relevant to the 2011 London riots. In the town of Enfield, police received reports of disorder an hour and 23 minutes after computer systems could have picked up the same information from Twitter, according to the study. "In this research, we show that online social media are becoming the go-to place to report observations of everyday occurrences -- including social disorder and terrestrial criminal activity," said co-author of the study Pete Burnap.
Light travels faster than meat.
Or 999 if you are in the UK. Just crowd source solutions to your problems.
. . . idiots tweet and post selfies when they should call the police instead.
How hard is it to detect a riot? Hey look, there's a guy throwing a trashcan through a window surrounded by dozens of other people. Doesn't seem like rocket science.
But totally useless for orbital or extraterrestrial criminal activity.
Yet again our Slashdot overlords subject us to clickbait and misleading titles.
This should read, "Twitter Theoretically Could Detect Riots Faster Than Police."
In other news, Jennifer Lawrence Could Theoretically Show Up At My Door And Demand Sex.
Get rid of white privilege and a lot of riots will go away. I'll get lots of negative replies because Slashdot readers tend to be bigots in denial. However, it's true that a lot of the civil unrest is due to racism that still exists. Slashdot readers tend to actively deny this because this site is basically the new Stormfront and the readers are threatened by minorities that might diversify technology. Racism and sexism still exist and, in my experience, most of you are racists and sexists. Get rid of the institutions of inequality that you help promote and perhaps the civil unrest you hope to detect, just won't exist any longer.
Last night I finished reading "Hatching Twitter: A True Story of Money, Power, Friendship, and Betrayal" by Nick Bilton. There's a quote by Mark Zuckerburg that the Twitter founders "drove a clown car that feel into a gold mine." They played musical chairs for the CEO for first few years and the current CEO is a Steve Jobs wannabe. Unbelievable.
Very few modern riots are spontaneous. Usually, they are orchestrated political affairs by those who wish to intimidate others. Follow the money. Twitter can show the planning. Be aware of the world.
"Liberalism is a very noble idea, currently controlled by some very bad people. Be sure you do not get the two confused.
Monitor Twitter and find the riot faster...
It will be all fun and games until somebody figures out how to spoof a riot by spamming Twitter... The police show up and volia! A new way to "SWAT" someone...
BTW.... For most riots... Who doesn't know in ADVANCE where they are going to be? We act like they are just events that happen at random times and places. You may not know the exact block the violence will break out, but it's usually pretty obvious when the risk of such behavior is high and where it's likely to happen based on the current events driving the whole thing.
Riots, like fire, have some pretty easily identified prerequisites.... 1. Groups of people gathering for some reason... 2. Strong emotions around the reason... 3. Strong rhetoric associated with the reason, encouraging people to feel hopeless about affecting some kind of change 4. A faction of people involved who don't mind violence.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Okay, they said the London riots were predictable from Twitter. Did they get additional positive correlation from other twitter --> actual riots? Did they attempt this and find that they couldn't get the correlation? Did they predict other riots that simply didn't materialize? I'd like to believe there's more to the study than is found in the story, but frankly there just isn't enough there to indicate a xx% rate. A data point of 1 is not a recipe for actionable data.
Until it's used to actually report the riot before the authorities I'm skeptical. It's easy to look back and see that there were clues that a riot was going on, but it's another thing to look at the present and say a riot IS going on. You also could get plenty of false posities. Not to mention the moment you use it to predict ONE riot the Trolls will figure out how to trip your detector.
If law enforcement doesn't already have back-end access to this data, they will soon. I imagine the information gets even more detailed and accurate once you aggregate data from all the tracking companies (Apple, Facebook, Google) together. Not to mention the NSA's own databases.
You could easily imagine a system being developed to track not just currently-happening riots, but the likelihood a riot will happen in the near future. Which area it's likely to happen in. Who is likely to participate.
Of course, charging people in open court with pre-crime probably won't fly. And a predictive system like this is always going to be a game of probabilities, it might not have the "resolution" to predict individual actions. What you can do is systematically target groups of people for surveillance or manipulation. Not that we aren't already doing that, but now it's more effective. And who the data tells us to watch might end up being somebody different than they watch today.
its great to look at how fast twitter "responds" to a real riot... now how many messages have been posted where once can interpret it as a reported riot where there isnt 1?
The ability to tweet and have it automatically picked up by the nearest emergency response org. seems like a needed feature.
and terrestrial criminal activity
I couldn't help but wonder if it could also find extraterrestrial activity.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
I was working in the bubs of Seattle during the WTO fun and games and I saw the substantial lag between events and police response. On local live TV I saw the black-block breaking windows and instigating the looting of Starbucks, etc. I also saw on live TV the protesters try and stop the damage/looting while the black-block made themselves scarce. A only after the fact were their police press conferences and riot squad response that only thumped on sweet young hippies in turtle costumes singing save the earth songs. I assume by the time the police hit the protesters, the black block was back at the bar watching the same live TV coverage I was and laughing their heads off. If just one cop would have been watching TV and radioed a couple of beat cops to grab a few of the right juvenile delinquents with the full help and support of the protesters the whole WTO riots would have been prevented. But since the city was crawling with feds/secret service just spoiling to have the local cops crack skulls the whole mess was inevitable. With lessons like these, the idea of some half-baked tarted-up search hit launching the riot squad at the wrong time/people/place does concern me. Good things there are no twitter bots out there to cause something to go wrong.
Can't wait until they send Hercule Poirot to solve "Murder on the ISS".
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Not that impressive. The Kaiser Chiefs, on the other hand, know before it even starts by using such techniques as measuring people's lairiness.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
could it also be used to predict impeachment ?
Nullius in verba
All very well to turn up the sensitivity to max and predict something. Not so easy to continue to predict things with the sensitivity turned down to the point where false positives don't flood the system.
Nullius in verba