Google Flutrends and predictive data mining

Google.org recently announced Flutrends, a web site that provides estimates of flu activity in the US based on search queries.

I'm not sure whether this is a good thing or just plain scary and I'm not surprised that they've found such a strong correlation between their data and the government's either. I am, however, quite interested in knowing what else is in there.

In Ian Ayres' book Super Crunchers: Why Thinking-by-Numbers Is the New Way to Be Smart it is mentioned that Wal-Mart use data mining techniques in order to match shelf space against predicted customer demand. Pretty cool. And they're not alone. But they use closed data sets. Their own, of course, and data that they buy from companies like Terabyte.

I hope that more companies and governments follow Google's example and open up (at least parts of) their data stores for public use. There are lots and lots of mash-ups available that show the potential of combining data sets and creating useful services to the public. Of course, for most companies they still have to find out that they actually have interesting data in their systems.