The internationalization of the 2014 West African Ebola Virus epidemic with case importation in Nigeria, Senegal and the USA has generated increased alert worldwide. This very dynamic situation is presenting policy makers, first responders and the wider public with a growing number of potential warnings and events (mostly unconfirmed) related to Ebola spreading and case importation. Following the multitude of event from regular media streams is proving to be more and more challenging. EbolaTracking relies on the Twitter global conversation to provide an awareness tool able to follow multiple events in a georeferenced context and in real time.
EbolaTracking taps into the Twitter Streaming API and monitors tweets mentioning ebola-related keywords. A machine learning system trained with the supervision of experts filters informative tweets. Geographical entities mentioned in tweets - such as country and city names - are identified using the GeoNames database and used to place tweets on a global map.
The map shows events, i.e., groups of tweets that recently mentioned the same place. Blobs indicate mentioned places (not the place those tweets were posted from). Blob size relates to the frequency of recent mentions. Countries are color-coded to indicate country mentions. You can click on a place to learn more about what is going on there.
EbolaTracking was designed and built by the Data Science Laboratory of the ISI Foundation (Marco Quaggiotto, André Panisson, Matteo Delfino, Ciro Cattuto) and by the MoBS laboratory of Northeastern University (Nicola Perra, Qian Zhang, Alessandro Vespignani).
