Among the first things that would likely come to your mind when you think about Twitter are trivial updates that range from what your friends had for lunch and how celebrities looked in an ongoing red carpet event and while Twitter can also be used as a source of important news, it is seldom associated with tracking diseases, which it actually can.
In a new study published in the journal PLoS ONE December 9, a group of researchers from Johns Hopkins and George Washington universities found that Twitter can accurately track flu trends at a local level. The researchers were able to gauge the spread of the disease in New York City by analyzing flu-related tweets from the area.
The researchers said that they wanted to analyze the performance of their influenza infection detection algorithm on a local level and deployed the system during the 2012-2013 influenza season. Describing the results of their study, the researchers said that their surveillance data correlated with the data from the Centers for Disease Control and Prevention (CDC) and Department of Health and Mental Hygiene of New York City.
The researchers also said that their system efficiently detected the direction of influenza prevalence. "Our system detected the weekly change in direction (increasing or decreasing) of influenza prevalence with 85% accuracy, a nearly twofold increase over a simpler model, demonstrating the utility of explicitly distinguishing infection tweets from other chatter," they reported.
Study lead researcher David Broniatowski, from the Department of Engineering Management and Systems Engineering of the George Washington University, said that their system worked in detecting flu trends on a national and local level. He also said that tracking localized flu trends is valuable because local flu activity may differ from national trends.
"Not only did our results track trends on the national level, but they also did so on the local level," Broniatowski said. "It gives our system validity. It shows that we're measuring what we say we're measuring, that we're tracking very useful information. And that localized data is valuable because the flu activity in, say, Boise, Idaho, may be quite different from the national flu trends."
