The flu is an endemic disease, but its prevalence increases substantially during what is commonly called the “flu season”, roughly between October and March in the United States (Northern hemisphere) and between March and October in Australia (Southern hemisphere). The Center for Disease Control and Prevention (CDC) uses a network of doctors that report influenza-like illnesses (ILI) as a share of their patients. This data is collected and provides a good indicator of flu activity in the United States. However, this data takes between one and two weeks to be collected and aggregated, implying a lapse in reporting.
Since most of the population of the United States has access to the Internet and that Google is the most used search engine, there is a high probability that once someone (or someone close such as a family member) develops flu-like symptoms, the information will be sought on the internet using a search engine. Thus, by monitoring and aggregating revealing search queries in Google (about 45 terms such as “cold remedy”) and comparing them with total searches, it is possible to monitor almost in real-time influenza prevalence. The above graph displays such searches between 2003 and 2015. The evidence so far underlines a close match between Google searches and ILI, so search engine prevalence and influenza prevalence are similar.
Comparing the United States and Australia, in addition, to reveal a counter-seasonality, also underlines a level of “over-concern” within the American population, which may be related to how the media covers the flu season.