Google Search Insights provides data on the volume of searches on selected terms, worldwide, updated daily. I've previously played with this data looking for evidence of an "Atlas Moment". Google has specifically played up the potential for this kind of nowcasting to help track flu trends. And this looks to be a bad one.
Let's go to Google Search Insights and see whether there was any increased search volume on influenza in Mexico as this developed. On a 12-month track, we indeed see a huge spike in flu searches, in Mexico, recently.
That's not particularly helpful though. Let's narrow it down to the last month.
We can see that things started to blip up around 22 April. A troll back through Google News shows stories starting to hit the papers around 23 April; one of these points out that Mexican doctors were already then well aware of the outbreak. Mexico issued its public health advisory on 22 April.
Perhaps the problem is that search insights scales to peak intensity and we ought here be using a log scale to avoid dampening out early warnings. Unfortunately, I can't narrow down the search to April prior to 22 April. But there's no upward trend in March.
I see no evidence that search intensity in Mexico increased in the leadup to the influenza outbreak. However, if the mess started in places where folks live in close proximity with swine and chickens then perhaps we shouldn't be too surprised that the first thing affected folks did wasn't a Google search.
The worldwide search trends are interesting. Mexico does have the highest search intensity, followed by Indonesia, the US, and Canada. Indonesia. Hmm.
Long story short, I can't see any way that folks could have predicted the flu outbreak by having carefully watched search intensity in the leadup to last Wednesday. It would be great if Google could allow some search date operators that could exclude recent traffic: I'd love to see what the chart would look like for April, in Mexico, until 22 April.
Update: NBR points me to a Google Map tracking worldwide cases.