Everybody's talking data-mining these days - but if you're just starting your data-mining project, you're too late. The time to benefit from a fad is before it begins, not after it's mainstream. Are you really going to discover some text-parsing algorithm that hasn't been found by the tens of thousands of PhDs working on this already? Just saying. Now there may be some interesting ways you can use this sort of information. But only as an aid - don't make it the centre of your business plan. It's too late for that.