Using patterns in Google searches to predict flu outbreaks
Marcus du Sautoy explores how internet searches are building up a huge database of linked information which can then be mined for patterns. Visiting Google鈥檚 offices, he sees this at work and reveals just how closely Google can predict flu outbreaks, even before hospitals have begun to need to respond to them. Flu kills hundreds of thousands each year, but has a good seasonal pattern. It is, however, difficult to know where and exactly when outbreaks will occur. By looking at which search terms matched flu activity, Google were able to build a model, based on search terms, which can predict flu outbreaks.
Duration:
This clip is from
More clips from The Code
-
Hexagons in the natural world
Duration: 05:13
-
Imaginary numbers at use in the real world
Duration: 04:32
-
Using probability functions to help locate serial killers
Duration: 04:33
-
Normal distribution in fish populations
Duration: 04:09