Interesting article posted in The Guardian (see below). Here's my answer to this question:

A data scientist (I like the term data wizard better) is someone who can consistently derive money out of data, e.g. working as an employee, consultant or in an other capacity, by providing value to clients or extracting value for himself, out of data. Even a guy who design statistical models for sport bets, and use his strategies for himself alone, is a data scientist. But a journalist who writes an article about big data, is not a data scientist. In my opinion no diploma is required to succeed as a data scientist: you can learn the principles online, for free. Rather than knowlege, what makes a data scientist successful is craftsmanship, intuition and vision, to compete with peers who share the same knowledge but lack these other skills. If you work with people, in addition you need to have great communication skills and be able to guess what clients or your boss want.

What is a data scientist? EMC2 graphic representation of their survey. Click the image to see it

Everybody loves a data scientist: ever since Google's Hal Varian told the world that

the sexy job in the next ten years will be statisticians. People think I'm joking, but who would've guessed that computer engineers would've been the sexy job of the 1990s?

That, combined with the McKinsey report into big data last year is a powerful blend. The report reckoned the US alone would need 190,000 deep analytical 'data scientists' - and another 1.5m data-savvy managers to make the big decisions.

This week, in California, over 2,000 people grappled with this issue at the Strata data confe.... Data is big business, with companies like Google, Facebook and LinkedIn - and possibly every other corporate body you've heard of - creating huge profits out of the way they use data. This is the 'big data' everyone's talking about - the 2.5 quintillion bytes of information created every day from our internet searches, purchases, mobile phone calls and social networking. (If you're interested, a quintillion is 1,000 times a quadrillion, which is 1,000 times a trillion, which is 1,000 times a billion).

Read the full article at http://www.guardian.co.uk/news/datablog/2012/mar/02/data-scientist