The annoying posts about the weather that flood your Twitter feed during every snowstorm may be more useful than you think.
Tweets like “It’s snowing like crazy out here!” or “Why haven’t they plowed my street already?” can be more than just irritating reminders about the annoyances of winter. It turns out that they can actually help make roads safer for drivers.
Researchers at the University at Buffalo looked at how these types of weather-related tweets could be analyzed to improve computer models used for suggesting safe-driving speeds and roads to avoid.
"It doesn't matter if someone tweets about how beautiful the snow is or if they're complaining about unplowed roads. Twitter users provide an unparalleled amount of hyperlocal data that we can use to improve our ability to direct traffic during snowstorms and adverse weather," said Adel Sadek, PhD, director of UB's Institute for Sustainable Transportation and Logistics, and the study's lead author.
SEE ALSO: You Don’t Need a PhD to Hunt for Black Holes
Since tweets typically come with GPS data, the comments that Twitter users make about road conditions and weather can add to the knowledge traffic planners gain from cameras and sensors used to create their models. Since the models don’t take into consideration lingering ice after a storm or the fact that some streets have been plowed, their accuracy is sometimes limited.
Using 360,000 tweets from the Buffalo Niagara region over the course of 19 days in December 2013, researchers pulled data from about 3,000 relevant tweets. They searched for keywords words like “snow,” and “melt” and divided the tweets into “weather utterances” and “weather reports.”
An example of a weather utterance would be:
The roads are a hot mess out in the burbs all over. Snowing like CRAZY up in here ... drive safe everyone
Whereas an example of a weather report would be:
#BuffaloNY #Weather #Outside. #Cold #Snowing #Windy. @Parkside Candy
If enough “weather events” were observed, then they were counted as “Twitter weather events,” and using geographic coordinates, researchers were able to map the exact weather locations.
They noticed an emerging pattern: As snow falls, weather-related tweets increase, and vehicle speed and traffic volumes decrease. By incorporating this data into the models, they were able to increase the model’s accuracy, however, Twitter data was found to be more effective during daytime hours and in highly populated areas.
Future research plans include collecting data for longer periods of time and in different locations. So next time road conditions are really bad, consider helping people out and tweeting about it.
You may also like: Solar Storm Forecasts Prevent Tech Damage
