Art of Datamining

I really don't spend much of my time doing datamining. (I spend it doing that neural and language stuff.) However, I've spent a bit of time on it and pay attention to it.
A key point is that datamining is not a science but an art.
You can't apply a particular technique in every case.
You need to get into the data, wallow around in it, and try to understand it.
There are lots of techniques that can help.
Someone still needs to figure it out, to get it to work.
That fits in really well with the no free lunch idea.
Someone has to look around at the data.
Of course, there is still a question about whether spending time wallowing around is worthwhile or not, but that depends on the data.