Monthly Archive

Baby girl Jones is doing fine, with no real updates to speak of, so I’m making this post in lieu of baby news. I know, I know. You’re probably already disappointed.

I remember the first time I flipped through The Visual Display of Quantitative Information by Edward Tufte. My initial skim of the book was a kind of coffee-table experience — I was captivated by the aesthetics of the graphics without taking the time to appreciate their informational value. After a more thorough reading during my graduate coursework, chartjunk, small multiples, and other theoretical and practical concepts began to sink in. I remember disagreeing with more than a few of Tufte’s claims (and I still do), but I was enamored with his academic dedication to technical communication. It inspires me still today. I’ve been wondering lately…could I learn something about my musical preferences by visualizing my music’s metadata?

Even though I sometimes worry about Apple’s stranglehold on digital music, iTunes is the best digital media application available. I’ve been a loyal iTunes user for the past three years. Katie and I share our Mac, but I am responsible for 98% of the music uploaded. At a minimum, she deserves an understated tip of the hat: my wife is a good sport when it comes to my music-listening/buying/downloading habits. Since 2006, I’ve been able to collect and organize my music in ways that stacks of Case Logic albums could never accommodate. On the one hand, I miss liner notes and inserts. On the other hand, I’d prefer to filter and sort data fields click-by-click anyday over flipping through plastic sleeves in a book.

Last month, I decided to delve deeper into my (and Katie’s) music library. I began with a loosely-defined purpose and one particular variable. I wanted to analyze my song aquisition habits since the beginning of 2007 by genre. In other words, how have my musical tastes changed over the past year and a half? Of course, genre is an extremely subjective way to categorize. For example, I draw a clear line of distinction with my mind and ears between R & B, Soul, and Funk. For example, if the average person were asked to sort Donny Hathaway, Jill Scott, Poets of Rhythm, Bo Boral, and Mary J. Blige into these two genres, their results would likely be different than mine. Some artists (e.g. Rufus Wainwright, The Avett Brothers, Beirut, Air France) are pretty darn difficult to force into one bucket, but they can’t be duplicated and put into two buckets or divided among multiple buckets. I keep reminding myself that it’s okay if the genres are subjective — I’m the only one interested in dissecting my library anyway.

In most of the cases where genre blurs the boundaries of visualization, I used the category Alternative & Punk as a bit of a catch-all. As any ontologist will attest, homogeneity is crucial to characteristics of division. If genre is a characteristic of song division, then a couple of my labels don’t fit the bill. As a category label, Soundtrack is problematic because it is not homogenous with the others. Finally, the category called Blanks (also not homogenous) consists of music that has not yet been assigned a genre label.

Here’s a snapshot of my music library in July of 2008. The full data set, or all the music I own, is about 10,100 songs. The pie chart below depicts songs by genre.

Music Library, by Genre

So, World music jumped 2,450%, from two songs in December 2006 to 51 songs in July 2008. The statistically-significant increases from January 2007 to July 2008 were:

* attributed mostly to Chatham County Line

Lounge and Metal were completely flat (no songs acquired) over the year-and-a-half period, while I only added one single Blues song (1%) and six Soundtrack tracks (3%). Increases in all the other categories ranged from 9% to 45%.

Here’s the breakdown of song acquisition by genre:

Music Acquisition Trend, by Genre

This exercise has me thinking about other variables that, when displayed visually, might reveal interesting trends or patterns. Play Count and Skip Count would really describe my listening habits, but there’s no data because I rarely play music in iTunes. I suppose I could start appending each song record in my library with My Rating, but tastes change overtime and it would be a full-time job assigning stars to every song I hear. Perhaps the next time I sort through my music, I’ll look at the gradual trend of acquiring songs and not entire albums during the last several years.

I’d certainly like to hear any ideas you may have about visualizing music collections and listening habits.