Study: Tweets during Hurricane Sandy accurately predicted storm damage

Researchers say social media activity can aid government officials and emergency responders with rapid damage assessments during disasters

The emergence of social media as a real-time gauge of what's happening in the physical and digital world could increasingly prove to be valuable asset in predicting the extent of damage during catastrophic weather events, according to a new study that analyzed Twitter activity as Superstorm Sandy swept through New Jersey in 2012.

The study, published Friday in Science Advances, examined how citizens used social media to improve situational awareness, disseminate emergency information and coordinate relief efforts.

Using Twitter data from news organizations and unaffiliated users nationwide, researchers tracked 52.55 million messages from 13.75 million unique users between October 15 and November 12, 2012. They then narrowed the total down to 9.7 million geo-located Tweets, focusing specifically on New York and New Jersey, the states hit hardest by the storm.

"At the peak of the disaster," the researchers write, "event-related keywords rank higher and activity increases with proximity."

The study next turned to mapping how the social media data correlated to actual damage on the ground. By tracking FEMA assistance grants and insurance claims, the researchers concluded that social media activity offered a remarkably accurate reflection of real-time damage during Hurricane Sandy.

The researchers say that government officials and emergency responders should explore how they can leverage social media platforms to rapidly assess disaster damage during future weather events.