UNIVERSITY OF ROCHESTER, NEW YORK, USA. An epidemiological algorithm can scan tweets and can predict if a person will get sick before any symptoms can occur. This innovation integrated in smartphones can certainly help people avoid entering public places with high incidence of flu. It can also provide preventive measures if you are at high risk of getting sick over the next few days.
Researchers at the University of Rochester analyzed 4.4 million tweets tagged with geolocation data for over a month with the use of a special algorithm. The team was able to map flu incidences around the city. The algorithm can tell the difference between non-literal terms such as “so sick of work today” and those that express flu symptoms. With the use of the GPS tagging, the researchers can see where the sick-feeling tweeters are circulating in the city and which tweeters are circulating near them.
Analyzing the results, the researches were able to predict when healthy tweeters were about to get sick eight days in advance. Their predictions yielded 90% accuracy. The system still cannot account all the people who got sick that it did not catch. Their research clearly shows the power of Big Data in positively influencing people’s lives. It is still not perfect but it absolutely opened doors to interesting and innovative public health applications.
|Organization||University of Rochester, New York, USA|
|Researcher||Researchers at University of Rochester|
|Field(s)||Epidemiology, Epidemiological Algorithm, Social Networking, Twitter, Flu, Smartphone, Mobile Technology, GPS|
|Further Information||Big Think|