I keep thinking I am the absolute KING of typos these days, but it is just my FF dictionary telling me that none of my words are actual French words. Except shocking. Shocking is fine.

You could assign each song a “cheerfulness” value (subjectively assigned from 0-5) perhaps, or a “popularity” value (based on number of weeks on radio charts), or any number of other values, and then look for patterns in which songs are played most often, where in the show they’re played, in which cities certain songs are skipped, etc. It’s the kind of data where the outliers would be the most interesting parts.
“What? They’ve never played ‘Poets’ in Seattle?” That kind of thing.

I can think of a bunch of other bands I’d love to do this with, too. Then you could aggregate! “Wow, songs played in Memphis are 75% more cheerful than songs played in Austin for all five bands we looked at.”

This is how pet projects turn into monsters, I know, but when it’s about music I can help feeding them.

And… yeah, I’m not sure how I would have done a chloropleth map with city data. You need areas to fill in for these things, so that doesn’t make much sense if you’re using cities. I might have tried to do it by county, but having it be two separate countries that divide things up and identify areas differently made that really hard.

I definitely do want to do cities, but I think markers on a googlemap will be a better fit. That, however, will require me to get latitude and longitude coordinates for the cities, and if I’m doing THAT, I maybe want to do it by actual venue. Like, they have played a lot in Toronto, but I’d be curious to watch them go from smaller clubs to, say, the Gardens. I could do a set of markers for each tour, and you could turn them on and off and watch them take over the world! Or, you know, Canada.

I also agree that there’s a ton of cool stuff to do with setlists, but that is step four or five of my crazy plan to visualize all this stuff.

I’m curious: was the availability of a useable map what made the decision to visualize location by state/province rather than city? Or was it a granularity issue?

It would be potentially fascinating to see set list data, particularly if you found a way to correlate that with region, world events, days of the week, or Gord Downie’s coffee consumption. :-)