University of Warwick researcher Tobias Preis giving a TED Talk

After being criticized for inaccuracy in its Flu Trends, Google Inc., has decided to combine ‘real’ big data with its search engine stats to try and come up with a better model.

Google Flu Trends was launched in 2009, and while it was lauded for its accurate results in the first couple of years, it later become clear the tool was wildly over-predicting the incidence of flu cases when compared to data from the Centers for Disease Control (CDC).

The problem with using search engine data only is that very often, people don’t have flu at all – but they do display flu-like symptoms, hence they search for things like “Do I have flu?”.

Searches like that can skew Google’s predictions, and so the Mountain View firm is now combining data from the CDC with its own to try and come up with a more accurate picture.

“For the 2014/2015 season, we’re launching a new Flu Trends model in the U.S. that—like many of the best performing methods in the literature—takes official CDC flu data into account as the flu season progresses,” said Google’s senior software engineer Christian Stefansen in a blog post.

Google adds it will publish the results of this new method in a “technical paper” soon, where it’ll hopefully be able to tell us if the new model is more succesful. However, we’re confident that it will be, if this research paper from The Royal Society is correct. University of Warwick researchers Tobias Preis and Helen Susannah Moat claim that by combining Google’s Flu Trends data with historic flu records, “the mean absolute error (MAE) of in-sample ‘nowcasts’ can be significantly reduced by 14.4 per cent, compared with a baseline model that uses historic data on flu levels only”.

We’re hopeful that Google is about to get much better at spotting flu trends from now on, but in any case it’s highly recommended you go and get your flu shot soon, just to be on the safe side.

In the meantime, here’s a look at the current state of play in the USA, according to Google:

photo credit: kenleewrites via photopin cc