In 2008, researchers from Google announced that they could predict outbreaks of the flu up to two weeks before the US authorities, by monitoring people’s Google search behaviour.
The algorithm tracked searches for flu symptoms and remedies, which would increase in the build-up to an outbreak. Flu is a serious disease that can cause up to half a million deaths each year – so Google flu trends caused a lot of excitement in the field when it emerged.
But for all the hype, it didn’t actually work. In the end, Google flu trends failed pretty badly. So what went wrong?
This month, we speak to public health experts and computer scientists to find out, and learn how the field of ‘Digital Epidemiology’ has moved on to successfully track disease outbreaks online.
We hear from Professor Guy Poppy and Dr Sian Thomas from the Food Standards Agency in the UK about a tool to track norovirus using tweets, and all the #vomit that entails.
And we speak to Professor Alessandro Vespignani from Northeastern University in the US about how Twitter can help to build virtual synthetic worlds where researchers can model the spread of flu.