Big Data Researchers support Google Flu Trends
Google Flu Trends is a service to estimate the number of people in a country who currently have the Flu. It does this by looking at the number of searches caried out in a time period, and using that as a basis to estimate the number of sufferers.
Google's Flu Trends webpage explains: "We have found a close relationship between how many people search for flu-related topics and how many people actually have flu symptoms. Of course, not every person who searches for 'flu' is actually sick, but a pattern emerges when all the flu-related search queries are added together."
In early 2013, the service came under fire for inaccuracy - a media storm cause a massive spike in flu-related searches. As a result, Google predicted double the number of flu cases that the US Centre for Disease Control (CDC) documented.
So how can we make use of data with such a high margin for error? Tobias Preis and Suzy Moat, both of the Behavioural Science group at WBS, have an answer. Instead of just using Google's potentially inaccurate data, or the CDC's data which suffers from a roughly 7-day lag time while statistics from around the US are compiled, we should combine them.
Tobias and Suzy have devised a 'nowcasting' model for estimating current flu levels by doing just that. The model incorporates both Google Flu Trends and official CDC figures and by doing so achieves an improvement on models based on either Google Flu Trends or the CDC figures alone. A paper discussing their findings was published in Royal Society Open Science.
You can read more on this story in Medical News Today, where Tobias and Suzy have explained the implications of this work.