Songs of Summer Jobs
by Rob Mitchum
This is part of our Clarify Data Stories series. Check out the big issues, minus the noise at www.spotify.com/clarify.
There are few more established seasonal American rituals than the summer job. To fill their break and their pockets for the school year, high school and college students take on temporary work and internships at the nation’s swimming pools, ice cream shops, and golf courses. But summer is also a time for jams, favorite songs for unwinding after work and on the weekend. Since a summer job has the benefits of inflating your bank account and keeping you out of trouble, it’s possible that those with jobs also listen to more positive, higher energy music. As such, the music listening habits of America’s youth may prove to be a useful index of the mood, and even the economic status of young people across the country.
Today, in aggregate, U.S. unemployment is low and still dropping, currently down to 4.9 percent after the “Great Recession” peak of 10 percent in late 2009. But youth unemployment (defined by the Department of Labor as ages 16–24), remains high, set at 10.2 percent in August. Zooming in to the state level reveals a broad spectrum of youth unemployment, from only 4.5 percent in North Dakota up to 17.4 percent in West Virginia in 2015.
To test our theory on jobs and music, we used characteristics automatically calculated by Spotify for every song in their system: energy, danceabilty, and valence (similar to mood, with a scale running from “negative” to “positive”). Each measure is determined based on a song’s non-lyrical content: beats per minute, rhythm regularity, dynamic range, and many more features.
We broke down the most popular songs for 16 to 24 year olds in each state “above index” — because the most popular songs are the same for pretty much every state (read: a whole lotta Drake), it’s more interesting to look at the songs each state listens to more than the national average. With that data, we could find the top “unique” song for each state, which you can explore below and listen to here. Independent of any economic correlations, the map gives you a sense of what each state’s individual “summer song” was for 2016.
2015 youth unemployment rates and top song over index for each state. Data: Spotify, U.S. Bureau of Labor Statistics. Graphic: Diego Garcia-Olano
When you look at this group of 51 (District of Columbia’s in there) top songs, and plot them by youth unemployment and Spotify’s musical characteristics, there are wisps of a trend. Energy and valence skew downward; in general, states with high youth unemployment listen to lower energy, sadder music. Danceability, however, stays flat — summer job or no, it appears 16 to 24 year olds want to get down in the summertime.
2015 youth unemployment rate per state vs. the energy of each state’s top song above index. Data: Spotify, U.S. Bureau of Labor Statistics.
2015 youth unemployment rate per state vs. the valence of each state’s top song above index. Data: Spotify, U.S. Bureau of Labor Statistics.
2015 youth unemployment rate per state vs. the danceability of each state’s top song above index. Data: Spotify, U.S. Bureau of Labor Statistics.
North Dakota, the state with the lowest youth unemployment rate, prefers a song with the very economically astute chorus of “The dollar decides how far you can go.” That it’s in a song called “Porn Star Dancing,” well…technically, it is about at least two kinds of jobs. The absolute highest energy song is also found in a low unemployment state, Montana (unemployment 7.2%). Little Big Town’s “Day Drinking,” might not sound like an endorsement of responsible employment. But it does mention “Everybody’s always waiting on a Friday,” and “Blame it on the work day,” targeting its temptations at those who spent their summer in an office.
Low-end anomalies are also interesting to dig into. Idaho, with its relatively low youth unemployment rate of 8.2%, listens to a Yo-Yo Ma performance of a Bach composition (“Unaccompanied Cello Suite №1 in G Major, BWV 1007: Prelude”) more than any other state. Unsurprisingly, it doesn’t rank very high on energy or danceability. Yet somehow even less danceable is Marconi Union’s “Weightless Part 1,” the preferred unique summer track of Delaware, youth unemployment rate 12.0%.
On the other end, West Virginia, which leads the country in youth unemployment, frustrates the overall conclusions somewhat. Their top track “Wifin You” by Chicago rapper Montana of 300, is downbeat but romantically positive in its own way, with a chorus of “Swear, while I be pipin’ you/Swear, I’m thinkin’ ‘bout wifin’ you.” Farther down West Virginia’s list are nu-metal mainstays such as Five Finger Death Punch, Staind, Nickelback, and Slipknot, suggesting that at some point, feelings about joblessness flip from sadness to anger.
As any good scientist would hope, expanding the pool of unique favorite songs for each state reveals similar trends. Once again, energy and valence drop as youth unemployment increases, with valence looking like the stronger effect of the two (the R squared is still only .185, in case you think we’re doing real science here).
2015 youth unemployment rate per state vs. the average energy of each state’s top 50 songs above index. Data: Spotify, U.S. Bureau of Labor Statistics.
2015 youth unemployment rate per state vs. the average valence of each state’s top 50 songs above index. Data: Spotify, U.S. Bureau of Labor Statistics.
It’s fun to look at some of the extremes for these top 50 song averages too. That high unemployment/high valence outlier is Nevada, where perhaps the presence of legalized gambling takes the edge off of unemployment. On the energy graph, the low energy despite low unemployment outlier also makes intuitive sense — it’s Hawaii, where a low 8.7% unemployment produces a mellow mood among its youth. Meanwhile, Ohio scores the highest average energy despite a youth unemployment rate in the upper half of all 51 states and districts, which can probably be explained with a single band: Blink-182, who occupy the top 7 spots and 17 of the top 30 for the state.
Examples such as these, where a single band hijacks a state’s uniquely preferred songs, demonstrate why this conclusion won’t be appearing in psychology journals anytime soon. But
it’s no accident that the fields of economics and psychiatry share the word “depression.” Since as far back as the Great Depression, researchers have found evidence of a link between unemployment and mental health issues, including mood disorders, stress, anxiety, and low self-esteem.
The connection is particularly pronounced among unemployed young adults, who a 2015 CDC study found were more than three times more likely to suffer depression compared to their employed peers. At a time when suicidal behavior can be predicted from social media data, it’s logical that musical preferences might also serve as a proxy for both economic and psychological indicators — a finger on the pulse of American youth, blaring from open windows after work.
Thanks to the Clarify data team at Spotify: Manish Nag, Ian Anderson, and Nathan Stein