The Drive: Your Tweets can improve travel information

It's no secret that a light dusting of snow can bring traffic on metro area highways to its knees, turning a 20-minute commute into an hourslong white-knuckle drive. But instead of sitting there and stewing about it, tweet it.

(Of course, you'll have a passenger compose your snarky rant or wait until you get off the road to complain about how bad the roads are.)

Aside from being cathartic, those 140-character tweets can provide valuable information.

When integrated with computer models, they can help traffic analysts recommend safe driving speeds and suggest which roads motorists should avoid when the weather turns foul.

That's the finding of researchers with the Institute for Sustainable Transportation and Logistics at the University at Buffalo in New York. They set out to learn how social media reports might complement traditional weather observations and improve travel information relayed to motorists, said Qing He, who co-authored the study that appeared in the October issue of the journal Transportation Research Record.

"Tweets happen in real time and provide good information," he said. "You get lots of information on social media and better understand the problems." That can tell an agency that "we need to get a plow to [location] because we see all these tweets."

Traffic analysts often rely on models that take data from cameras and sensors embedded in the road, and from weather stations that relay information like temperature, wind speed and precipitation.

While an effective way to gauge the effect of inclement weather on traffic, the researchers say the approach doesn't provide specific information on road surface conditions. The models don't account for icy conditions that may still persist after a storm had subsided or factor in vehicle speeds on roads that have been properly plowed.

Twitter can fill that gap. Users — who send more than 6,000 tweets every second and often reveal their locations via GPS — file on-the-spot reports that can be fed into traditional models and give a more complete look at road conditions.

For example, during the Twin Cities' first snowfall this season, Drive reader Jonathan Olson tweeted, "NB 35E from Burnsville split to DT STP 45 min, avg speed 40. No plow or sand/salt at that time."

The researchers in Buffalo, one of the snowiest cities in America, examined more than 360,000 similar tweets over a 19-day period in 2013. They spotted 3,000 relevant tweets by tagging words such as "snow" and "melt" and found that when snow falls, weather-related tweets went up, vehicle speeds dropped and traffic volume slowly decreased.

"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, the lead author.

That's key because nearly one in four crashes — 1.3 million of them — is attributed to weather, according to data from the Federal Highway Administration. They led to 6,250 deaths and more than 480,000 injuries annually. More precise models can point agencies to roads that need to be cleared and recommend safe driving speeds, the research said. That can make winter driving safer for all of us.