Cool idea, but I'm going to quibble a bit on the Audience vs. Band Chart. Suppose for simplicity that a band has four songs A, B, C, and D that respectively 40% 20% 20% and 20% of people would prefer to hear at a concert. Now suppose the band can only play one song.

Which song should they play?

I don't think the answer to that question is a random mix of the songs with 40% song A, I think the answer is 100% song A. If you play any of the other songs, you make 20% of the audience happy, if you play A you make 40% happy, it strictly dominates. In other words, I don't think playing the songs at a concert in proportion to their last.fm statistics makes sense since at home listeners have as much time as they want but concert time is a limited resource that should be used on the best songs.

So I would be interested in the rank order of listening stats vs. rank order of concert songs and look for anomalies. Unforgiven (I, II, III?) are probably going to make that list since I've actually heard of that one and I am definitely not a Metallica fan.