You Can't Always Trust Mash-Up Data

MashedThere's nothing we like better than a good mash-up (here's a couple of random examples). But when it comes to services that integrate data from multiple locations and draw their own conclusions, it's always best to apply a little healthy cynicism.

Picture by santos

In a session at the Gartner Symposium in Sydney today, Gartner analyst Andrea Di Maio pointed out that while governments and other organisations making data available for use in web applications was a good idea, it was dangerous to assume that the conclusions drawn from mashing together different data elements were as robust as the original data:

The accuracy of mashed-up data is going to be an issue. Those who mash up data always have a vested interest. People often want to give a specific biased view of some data.

Di Maio quoted a maxim used by a colleague: "If you torture the data enough, it will confess anything. There is this almost religious belief that the data is only going to do good, and I don't think that's the case," he said.

That doesn't mean that services that track swine flu or bushfires are without value; far from it. But it does mean that -- just like everything you encounter online -- you need to question the motives of the people involved and look carefully at the source information.


Comments

Be the first to comment on this story!