Health IT

Is Twitter a better predictor of the flu than Google and the CDC? Researcher says, yes


[Photo Credit: Adam Sadilek, University of Rochester]

Everyone in the U.S. who hasn’t been living under a rock is aware that we have been hit by a particularly pernicious flu season.

And the key to controlling the influenza or any contagion is not only early detection, but also in predicting the trajectory of the disease.

Here, social media, especially Twitter, is playing a big role.


Adam Sadilek, a researcher with professor Henry Kautz in the computer science department at the University of Rochester, has done an interesting and rather extraordinary experiment.

Using Twitter and geo-tagging, Sadilek has modeled how an infectious disease may spread in real time by plotting data on a map using tweets from people and seeing how healthy people and sick people interact. That led them to create a heat map of New York.

Sadilek asserts that this shows “emergent aggregate patterns in real time with second-by-second resolution,” which apparently suffers none of the time lag that other, more established models coming from the likes of Google Flu Trends and the Centers for Diseases Control and Prevention.

You can see a video of how the flu overtook the Big Apple over one day here: