Cook et al (2011): Assessing Google Flu Trends Performance in the United States during the 2009 Influenza Virus A (H1N1) Pandemic

Google FluTrends (GFT) is an algorithm that uses historic search-term flu-breakout correlations to provide near-real time estimates of influenza activity. The original model provided very good results for seasonal outbreaks over several years, but in case of the 2009 H1N1 (swine flu) pandemic, it overestimated the magnitude of the outbreak by a factor of two. In this paper, researchers from Goggle argue that Internet search behavior changed during this outbreak. The complications associated with swine flu, the fact that the pandemic began in the summer rather than winter, and changes in health-seeking behavior each may have played a part. An improved algorithm is presented.