Google released Google Flu Trends today, which harnesses the geographical distribution of flu-related searches to come up with a real-time picture of flu activity a good two weeks earlier than systems relying on CDC surveillance data. Hitherto, by the time the CDC data were published warning of high flu activity it most likely was too late to take any steps to protect yourself (i.e. wear face masks). (This scary graph shows an upcoming jump in flu cases. We’re tracking for imminent doom.)
Google’s Flu Trend graph is remarkably similar to surveillance data garnered by the CDC:
Flu trends is a very interesting concept, not only because it improves on current surveillance systems, but also because it’s not relying on conscious user participation. In an alternative model drawing upon social participation, people report in a flu tracker that they’re sick; those reports are then mapped. An example of such a tool is Whoissick.org.
Such systems are hard to implement in a way that provides accurate results, because in order for them to work, a vastly higher percentage of people would need to participate than currently participate in any Web 2.0 activity. Especially when you’re sick, the noble act of alerting other people may escape your fogged up brain.
But searching for health information related to flu symptoms is already natural user behavior. Google’s Flu Trend and doesn’t require any additional conscious effort on the part of the sick crowd.
What other ways could search query trends be harnessed to provide similar services based on the unconscious wisdom of the crowd, in the health domain or elsewhere?