Rather than post songs by the same artists I’ve been discussing over the past seven weeks, I’m going to present each day’s findings alongside a YouTube video of a cute girl covering a Bob Dylan song.
I found a bunch of these clips earlier in the year, posted a few of them and stashed the rest for a rainy day. This week it’s raining genome data.
The first step in preparing my research was to rank the 35 artists I included in the study in order of preference. Rather than create a strict 1-35 list, I split the artists into five groups of seven apiece. This was the result (names within each tier are listed alphabetically):
Tier 1: Fiona Apple, Belle & Sebastian, Elvis Costello, Bob Dylan, Ben Folds, Paul Simon, Rufus Wainwright
Tier 2: The Beatles, Aimee Mann, Tift Merritt, Randy Newman, R.E.M., Ron Sexsmith, Lucinda Williams
Tier 3: Miranda Lambert, Lyle Lovett, Tom Petty, Josh Rouse, The Shins, Elliott Smith, Bruce Springsteen
Tier 4: Counting Crows, Fountains of Wayne, Billy Joel, John Mayer, Brad Paisley, Michael Penn, The Rolling Stones
Tier 5: Beck, Eminem, Shakira, The Smiths, Stew, Talking Heads, Dar Williams
All of the usual list-making caveats apply. I could listen to an album by a Tier 3 artist tomorrow and think I was nuts not to put him or her in Tier 2. But this is the arrangement I’m using for the purposes of this research.
The next step, and the one I’ll focus on for today’s post, was to see how many categories each of these artists fell into, and to look for any patterns across the tiers.
My hypothesis was that the more I like an artist (ie., the higher a tier she is in), the more genome categories she will fall into. Each tier, therefore, should have a higher ‘categories per artist’ number than the one below it.
Hypothesis confirmed, more or less.
Here are the average categories per artist for each of my five tiers:
Tier 1: 2.86
Tier 2: 2.86
Tier 3: 1.43
Tier 4: 1.43
Tier 4: 1.00
As you can see, the more I like an artist, the more likely he or she is to show up in multiple genome categories. Three of my four oddball, zero-category artists wound up in the bottom tier, while the top two tiers were dominated by those in two, three or four categories.
This is one way to answer the question that came up several times over the past seven weeks (“If you like ‘pop’ or ‘folk rock,’ why don’t you like so-and-so?”). Maybe so-and-so is just a pop or folk rock artist, while I tend to prefer artists who combine those sounds with my other categories.
So my perfect match would be a melancholy singer-songwriter who performs in folk, pop and country styles and plays the piano. Randy Newman was the closest of my 35 candidates to that ideal, though the ‘pop’ and ‘country’ designations are a bit of a stretch.
Actually, the artist who comes to mind when I read that description is Elton John. And that might be an argument for me to seek out more of his albums. I own only a few, and don’t listen to those very often — not because I don’t like them but because I’ve never taken much time to get to know them. Maybe that should be my first takeaway from the genome project.