For a better browsing experience, please upgrade your browser.

FiveThirtyEight

Science

Yesterday, I had the opportunity to speak with a couple of engineers from Google who are working on a product known as Google Flu Trends. This is a very simple, yet elegant and important application of what might be termed predictive analytics; if there were awards given out for such things (the Jameys?), it would be a good candidate to win one.

The product, which launched last year, works by analyzing searches that have correlated strongly in the past with flu statistics as put out by the CDC and other governmental agencies; a fuller write-up of the technology can be found in this article in Nature. The advantage of this is that whereas the CDC typically works on a 10 to 14 day lag before new flu statistics are published, the Flu Trends numbers can be turned around literally overnight. Flu Trends does not predict the future per se, so much as it “predicts the present”, as the engineers describe it.

The other nice thing about the Flu Trends data is that it is all publicly available. Here, for instance, is when the Flu Trends index hit 5,000 in each U.S. State, a level that would correspond to the peak of a fairly bad annual flu outbreak in the January or February.

This map is fascinating on a number of levels. Although the initial outbreak of H1N1 back in April was centered on Texas, California, New York, Illinois and South Carolina, the place where the flu first hit critical mass several months later was in Louisiana. It then slowly radiated its way outward to most of the neighboring states — Maine finally hit the 5,000-point threshold just last week. There also appear to be other points from which the flu spread — a less prominent ‘epicenter’, for instance, centered in Minnesota and the Dakotas. And somehow, there came to be quite a lot of flu at various points in both Alaska and Hawaii — Hawaii’s peak actually came way back in June and July, well before the one in the Deep South.

The flu has not been especially widespread in Florida, perhaps because Florida has a lot of old people and — unlike the seasonal flu — H1N1 has mostly attacked younger individuals. The other state which has yet to hit the 5,000-point barrier is Utah, which is somewhat culturally isolated from the rest of the country.

The good news is that, according to Flu Trends, the flu is pretty much on the decline in all states except Northern New England. The bad news is that we’re starting to approach the point where the seasonal flu usually starts to grow in advance of its typical January or February peak; how these two trends will intersect, I don’t purport to know.

Filed under , ,

Add Comment

Powered by WordPress.com VIP