31
2005
Avian Flu – Hitwise Data in The Economist
I just returned from my visit to the UK. On the flight over, I picked up a copy of the Economist at the Airport newsstand and was thrilled to see Hitwise data cited in their Avian Influenza – In a Flap article (October 22, 2005 page 82-84). The chart in the Economist shows the combined searches for “bird flu” and “avian flu” in the U.S.
Below is a chart that didn’t make the article where I separated out the two terms.
Separating terms yields some interesting data. You’ll notice that there was a small spike in searches in early April on “bird flu” that occurred after the spike in “avian flu” in early March. Using Hitwise Search Term intelligence, I noticed that there was a distinct difference in sites receiving traffic from the term. “Avian flu” was sending traffic to informational sites such as www.cdc.gov, while “bird flu” was sending traffic to news sites such as news.google.com. Hypothesis: Increase in searches in April was probably the result of a specific news story on the subject. But why did this specific spike not lead to visits to informational sites (as it had on other occasions)? Any guesses?
Hi Bill
BTW interesting site you have. I’ve long reported on ‘avian flu,’ as a journalist, so this caught my eye.
You are quite right to have thought of separating out the terms “avian flu” and “bird flu” as these do generally resolve different demographic populations of both searchers and newspaper articles; you could take this a step further, and test, as well, “H5N1″ and “avian influenza”; the last term will almost entirely represent searches by scientists and health professionals, so it could be interesting to see its distribution, and that of H5N1, compared with “bird flu and avian flu”.
The April peak is interesting. As well as reporting on avian flu, I share my journalistic reading/research as a daily blog/newswire. As far as I can remember, internationally, there was nothing in April that I can see that was exceptional in terms of newsworthiness, that could have produced such a pronounced peak (late May was much more active in events); do you have geographical data on terms, to see if this peak was limited to an event or a story in one country or region?
Re: “But why did this specific spike not lead to visits to informational sites (as it had on other occasions)?” One hypothesis would be that this had a single/local point origin as you suggest, in which case something like a Zipf power law might suggest that recommended links, in this case from a few news sites, just took the lion’s share of the the links, in which case the pattern would not be that surprising?
Anyway, just some rough thoughts of the top of my head. Would be good to compare this data with data on the same search terms in news articles worldwide over the same period; data that Google News and other would have.
Hope these few quick thoughts help.
Declan Butler, Ph.,D
Senior reporter, Nature
7 rue Guy de la Brosse, 75005
Paris, France
d.butler@nature-france.com or d.butler@nature.com