Twitter app tracks flu outbreaks in real time

What people in your city are Tweeting could help you predict your odds of coming down with the flu … or avoiding areas where it seems that lots of people are getting sick.

A combination of machine learning and natural language understanding strategies can help us use computers to analyze the content of Tweets from any given geographical area, according to Adam Sadilek, a computer scientist at the University of Rochester. Such an application can identify hot spots where lots of people are Tweeting things like, “Home sick today” or “Think I’m going to throw up.”

Sadilek and his team have rolled out sample heatmaps for a half-dozen cities — including New York, London and Cape Town — that use the app to show the distribution of Tweets likely to indicate a case of the flu (such as, “Think I’m getting sick :(“) and other Tweets featuring the word “sick” that are less likely to be flu-related (for example, “Party last night was sick”). The most flu-suspicious Tweets are tagged in red.

“The more red an area is, the more people are afflicted by flu at that location,” Sadilek’s website explains. “We show emergent aggregate patterns in real-time, with second-by-second resolution. By contrast, previous state-of-the-art methods (including Google Flu Trends and government data) entail time lags from days to years.”

That real-time capability could prove invaluable in tracking flu outbreaks and other epidemics, even if the app isn’t entirely “smart” yet (for instance, it red-tags some non-flu-related Tweets like, “Damn I wulda been sick if I woke up n had a ticket”).

“Since a large fraction of (T)weets is geo-tagged, we can plot them on a map, and observe how sick and healthy people interact,” Sadilek’s website states. “Our model then predicts if and when an individual will fall ill with high accuracy, thereby improving our understanding of the emergence of global epidemics from people’s day-to-day interactions.”