The Gong Show

Premise
Two nights ago I started building a list of “best albums of 2006″ lists. I thought the aggregate of the data would be interesting to capture. I find reading best albums lists fascinating because I can completely agree with one choice and be totally turned-off by the next. I found myself reading so many of these lists, that I thought I should do a more formal deep dive.

Sources
I ended up compiling a list of 180 “top albums of 2006″ lists. Some are from old media: (think: SPIN). Some are from new media: (think: Pitchfork). Some are just bloggers. Some aren’t even bloggers, but left a “best albums of 2006″ list in the comments section of a blog.

My sources for finding lists were del.icio.us, technorati, my own OPML, music mags online that I like, and this excellent blog post I stumbled upon from somewhere.

I do not by any means claim this list to be complete or even comprehensive. For example, I didn’t go check out Rolling Stone online, that would be an important data point, but I got tired.

However, I do think I have enough lists to be interesting. I don’t know how many lists I would need to be statistically representative of the music community online; if someone with a stats background has the answer I would be curious.

Methodology
I did not descriminate based on readership, domain, nationality, or language. However, the content did have to be online (I didn’t feel like buying any magazines).

I only collected lists that ranked their lists in order of preference. Top 10 lists that didn’t have any numbers next to the entries didn’t quality because they didn’t provide enough information to do anything fun with the aggregate data.

I only collected lists that rated the best albums of 2006 without qualifications. For example, lists that said stuff like “the best hip hop albums of 2006″ don’t qualify because it is limited to a specific genre of music.

I only collected lists that mentioned bands and albums that were actually from 2006. Very few lists were removed based on this qualification, but some did get the axe. For example, one person on live journal listed the OK Go album as their favorite of 2006. This was actually released in 2005 so it didn’t qualify. I’m sure there are albums on this list that are not from 2006, and if you identify any please let me know and I’ll remove them. I only removed the ones I knew violated the rules.

Once I had my lists, I thought the obvious cool thing to do first would be to find out what is the general concensus on the best album of 2006. I had 180 sources to work with, so I simply copied the #1 album (and artist) on every list by hand into an Excel doc. A few pivots later I had my results.

Results
The following table shows the number of times a given album appeared in the #1 spot on a “Best Album of 2006″ list:

And here’s the long tail histogram of the table (labels removed for x-axis for readibility, but obviously it’s the albums in order of their rank):

Conclusions
Well, duh, the first conclusion to draw is that if a whole bunch of critics and bloggers were given one vote each, they would choose TV on the Radio’s Return to Cookie Mountain to be the best album of 2006. However, Return to Cookie Mountain represents less than 8% of the vote. Chris Anderson would be happy.

The second conclusion is rather anecdotal (though one could prove it with some analysis and more screen scraping): the aggregate of everyone’s number 1 choice looks very similar to people’s individual lists. If I get motivated over the break, I’d like to formally prove the correlation I’m describing, but after looking at 180 “best albums” lists I can say with confidence that taking only the number #1 pick from everyone’s lists is very representative of the average individual’s top 10 albums.

There’s a ton more that can be done with this stuff…. For example, if I did the same analysis on everyone’s #2 pick, would TV on the Radio come out on top again? Probably not because 8% of the lists would already have it picked in the #1 spot. Would The Knife be most picked at #2 (which would be a step towards proving my second conclusion)? Again, I’m skeptical, because ~5% of the people would have already selected it at #1…

Also, I think this table can be used for recommendations. If an album got at least 3 votes at the #1 album of the year, and you have not listened to it yet, then that’s probably something worth paying attention to. Indie geeks who think their tastes aren’t represented by mainstream culture might be turned off by that notion. So what. I know I enjoyed all the albums in the top 10 rank on this table that I have listened to, so I’m super excited to check out the ones I have not heard.

In closing, I should note that my own top albums of 2006 list was not included in this post because I never made one. However, if I did make one, I think I would put Girl Talk’s Night Ripper on top.

Finally, some humor to reward you for reading all the way down here.

References
My references (ie the links to all the “best albums of 2006″ lists) are all in the excel doc I made, which is yours to play with if you desire. If you do extend it in anyway, don’t keep it to yourself. Show me what you’re doing just cause I’m curious.