Scientists from the Los Alamos National Laboratory decided to take a look at Wikipedia search data in order to see if they could predict outbreaks of infectious diseases.
They chose to focus on seven diseases (Ebola, cholera, dengue fever, HIV/AIDS, influenza, plague, and tuberculosis) in nine countries (Brazil, China, the United States, Haiti, Japan, Norway, Thailand, Poland, and Uganda), and came up with models for 14 "location-disease contexts," the Los Angeles Times reports. Wikipedia collects data on the millions of search requests it gets daily, and researchers used data from March 2010 to February 2014. They also used official disease incidence reports to determine if the searches predicted an outbreak.
They found this worked for influenza and dengue fever, both seasonal diseases with short incubation periods that people anticipate. They were not as successful with the plague in the United States, since the number of people infected is so low, or Ebola in Uganda, where internet access is limited and those outside of the country were also searching for information on the disease.
Writing in the journal PLOS Computational Biology, the researchers stated that this Wikipedia-based approach "is sufficiently promising to explore in more detail."