Listening Habits + Playlists

Since 2006, I’ve been making what I call “Seasonal Playlists” for myself every few months. I have found that it is a great way to keep track of songs that I become obsessed with during a period of time, and these playlists have become an invaluable resource in terms of tracking how my musical tastes have changed over time. Including a specific song on a seasonal playlist is the ultimate sign that it means a lot to me at that point in time, and I want to remember it for years to come.

A couple of years ago, the first analytics post I wrote for this blog revolved around the idea of using Paul Lamere’s “Sort Your Music” tool to take a look at the musical characteristics of the songs in each of these playlists – such as average BPM, valence/mood, release date, popularity, and so much more. This time around, I’m focusing more on my listening habits as it relates to my playlists.

All the visualizations below were created using Tableau, along with my own personal scrobble data loaded into a CSV + some data scrubbing in Excel.

Seasonal Playlist History (Spring 2017 – Spring 2019)

I often find myself going back to my seasonal playlists when I don’t have anything to listen to, and they provide a great jumping off point to re-discover some of my favorite songs, artists, and albums from years past. That got me thinking – am I too reliant on my old playlists? Do I go back to a specific playlist or set of playlists more often than others?

playlists over time lines

Each colored line on the above line chart represents a running sum of plays over time for the songs in a seasonal playlist, going back to January 2017. As an example, the orange line represents my Summer 2018 playlist. The orange shading behind it represents the time period when I was creating this playlist (6/1/18-8/31/18). As I listened to music, the number of times that I listened to the 21 songs that eventually ended up being included on my playlist for that period increased, causing the line’s slope to increase as a result.

What strikes me the most about this visualization is the consistency of how I listen to the music that I end up including on my playlists. I can see that I will typically listen to the songs that end up making the playlist about 300-350 total times while I’m building it, before cutting myself off at the beginning of a new playlist period. I’ll call that cut off point the “inflection point,” and without fail, it occurs late in each colored area. That is the point in time where I’m typically satisfied with the songs included and the running order of the playlist. After the inflection point, I’m not listening to those songs as much, therefore the running sum does not accelerate nearly as fast, and the result is a much flatter slope.

These seven lines represent the seven playlists that I have made since I have started obsessively tracking my listening history via Last.FM in 2017, so it is not unusual to see so much of my scrobbling history for the past two plus years in this chart. But what if we go back further?

Wow, that’s a lot of playlists

Seasonals 2013 to 2016

Here’s a similar chart, this time each line represents the running sum of plays from 2017 on for playlists that I made from the Summer of 2013 through the Fall of 2016. The darker lines are the newer playlists from that time range, and vice-versa. These darker lines are at the top of the heap in terms of the number of listens (the top five lines are: Fall 2016, Summer 2016, Fall 2015, Spring 2016, Summer 2015 – in that order), and you can see a little bit of separation between these top five lines and the other six lines representing playlists from 2013 and 2014. I would imagine that over time, the separation between the later playlists and the earlier playlists in this set will continue to grow.

College Mixes

Finally, here are my college playlists (Fall 2009 – Spring 2013) with the bluer colors representing the ones I made during my senior year. They are far and away the leaders here, and by comparison, I don’t go back to the songs from my freshman and sophomore years as much. But, notice the y-axis – we are now dealing with significantly smaller amounts of listening.

Girl Power

Last summer, I took note of the amount of female fronted indie and rock groups I was enjoying such as Hatchie, Alvvays, Snail Mail, and King Princess. That statement and my perceived listening habits made me curious: am I listening to more female fronted bands recently? And, if that’s true, how do my changing listening habits impact my playlists?

The first thing to do to answer these questions is to find out whether or not my hypothesis about listening to a higher percentage of female fronted bands more recently was true. For that, I looked to my Last.FM scrobble data over the past 2+ years.

M vs F weekly

The chart above splits my scrobbles into two distinct groups, male and female, with points weekly, visualized here as a percentage ratio. I made the determination that an artist would fall into the “female” group if a) the majority of their songs were sung by a female, or b) in the case that the music was entirely instrumental, if a significant portion of the music was being performed by female musicians. That led to some tough calls: a band like the Pixies with Kim Deal gets placed into the male grouping, while Khraungbin – a mainly instrumental group – gets placed into female because of the strong presence of Laura Lee. I didn’t find any sort of data set that classifies artists in this manner, so this is not a perfect system, but it will do. To simplify the amount of manual classification and grouping, I made a cut off that I needed to listen to an artists’ songs at least 10 times from 2017 onwards. That cuts the data set a bit, but saved me a considerable amount of time, so it was worth it.

The biggest takeaway here is that the vast majority of music that I listen to is male (represented by the blue line). That isn’t surprising considering that my top played artists are The Beatles and Green Day. In fact, there have only been two weeks since 2017 where the majority of music that I listened to was by female artists, and both of those weeks occurred in the last six months. Even though growth has been slow, I have been listening to a higher overall proportion of female bands more recently.

The next step is to turn to my playlists to see what percentage of the tracks in each one were by female bands.

percent F in playlist revised

The results are indisputable here – I’ve been including a much higher ratio of female bands in my playlists, especially over the past year. This past spring, nearly half of the 23 tracks were by female artists, and the last three have been at least 33% female. In most cases over the past decade, my playlists have been packed with male artists, so this change is certainly significant.

The final thing to do was to put these data sets together to determine if there is any sort of correlation between the amount of female groups I listen to and the M/F ratio of songs in my playlists.

M vs F Bars and Lines

I grouped my scrobbles into the same time periods that I use for the seasonal playlists, and visualized my listening ratio using a stacked bar chart to 100%, with blue representing male scrobbles, and pink for female. I then overlaid the % of female artists in my playlists as a line chart over it. The results are a little blurry to tell on first glance, but the advanced statistics don’t lie – the correlation coefficient is .6919 and the R squared is .4787, meaning that there is some (but not an overwhelming amount) of correlation between my M/F listening habits and the M/F ratio on my playlists.

Thanks for reading! If you’re interested in more of this kind of music + data analysis, check out my series of analytics posts, and if you have any questions or comments, please reach out! You can find out more about what kind of music I listen to on my Last.FM profile, and if you’re curious about my seasonal playlists, you can check them out on my Spotify profile.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s