Twitter might seem like an enormous stream of aimless chatter. But, to an increasing number of agencies, researchers, and managers, it’s a valuable human sensing network, full of potential insights. We’ve written before about tweet-mining for disaster response (where social media is used to locate injured people and damaged infrastructure) and flu outbreak mapping. Here are another three uses: tweet-mining for emergencies, "social weather," and crowd management.
Richard Stronkman started Twitcident while he was a student at Delft University of Technology. His initial focus was helping emergency services in crisis situations, such as fires, terrorism, weather events, and so on. By filtering out useful messages from many thousands of noisy ones, the system can help managers decide where to put resources and manpower. Here’s Stronkman talking about a fire incident in Texas:
It’s one thing to have a weather forecast. But what actually happened when the weather came, how did it affect operations? The Dutch National Railway is using Twitcident to map weather, exploiting the Dutch habit of constantly discussing the weather." They have the forecasts," Stronkman says. "But it’s always difficult to know the impact. When the snow falls, for instance, will it stay on the ground? Because, if it does, a person needs to go to that location. This is eyes and ears on the ground that can help them."
Last year, Dutch police used Twitcident during busy events like New Year’s Eve, 3FM Serious Request (a charity music event), and Summer Carnival, when they knew overcrowding could lead to problems. "Whenever there’s overcrowding in a location, and people start tweeting 'it’s getting very busy here,' we try to filter the signals so the police know there is an increased risk. Then they can take action," Stronkman explains. At the Carnival event, in Rotterdam, police managed to quell rumors of a shooting (and ease a growing panic) by monitoring online comments and putting out their own messages. Two officers discuss the incident here.