2020 Has Changed the Way You Listen To Music — Here’s How

Image for post
Image for post
Oscillations in Americans’ musical tastes (and moods), from Tableau Public

The moments in my life which have impacted me the most are all colored by music. Songs by The Smiths are woven into the fabric of my childhood bedroom, the summers in NYC filled with a fresh smell and Biggie songs, and stressful evenings relieved by the soothing humming of Kid Cudi. 2020 has certainly been no different. I am more aware than ever now of changes bubbling underneath the surface, heightened or brought about by the hectic slew of events which this year has brought. My musical tastes have changed this year, of course, as they’ve changed every year of my life — they’ve changed to fit better to the contour of who I am now.

What makes this different is how universal our experiences have been in 2020. There have been many, many events which have affected all of us at once — protests, election coverage, and, of course, the pandemic. In a time which has made isolation and alienation more visible than ever before, we can turn to the music which we’ve listened to see how these universal experiences have affected us.

While in the past we might have had to rely on anecdotes or patchy records, Spotify’s data-driven approach to music can allow us to look at how Americans’ musical tastes have collectively changed with the ebb and flow of the past year. If we combine this with some cool findings from musical researchers connecting music to personality, then we can see how Americans’ personalities changed in response to current events, based on the most popular songs played in the US, on Spotify, on those days.

What Music Says About Your Personality

In “The Song Is You: Preferences for Musical Attribute Dimensions Reflect Personality”, musical researcher David Greenberg and his team found how the kinds of music which people listen to are correlated with various personality traits.

In order to objectively “measure” music, Greenberg and his team turned away from the idea of using genre to measure music. After all, what I think of when I hear “rock” probably isn’t what another person thinks of when they think of “rock.”

Image for post
Image for post
Your rock isn’t my rock. From top to bottom, left to right: David Bowie, Chuck Berry, and Iron Maiden- all rock.

Subgenres can help somewhat, but are often flouted by the same ambiguity which genres is general are. Is Travis Scott, for example, mumble rap, cloud rap, gangsta rap, or psychedelic? Instead of using genres or subgenres to describe music, the researchers mapped all songs onto two axes- arousal (or energy) and valence.

Image for post
Image for post

Songs which are high in arousal are very upbeat and tend to have characteristics like a fast tempo or low volume. Songs which are low in arousal might be slower or softer. Songs which score high on valence are very positive songs, whereas songs which are low on valence are more negative in mood.

Image for post
Image for post

Using these two metrics, Greenberg and his team then studied how they correlated with the Big Five, or OCEAN, personality model. According to the Big Five model, personalities can be described as a combination of five metrics- openness, conscientiousness, extroversion, agreeableness, and neuroticism (or sensitivity to stress). The researchers found that the traits are correlated with musical preferences as follows:

Image for post
Image for post

Red indicates negative correlation and green indicates positive correlation. So for openness, for example, as arousal increases, openness decreases.

As musical services such as Spotify adopt arousal and valence to measure music, Greenberg’s findings become increasingly filled with potential.

Getting Mood Metrics for the Most Popular Spotify Songs

Inspired by research such as Greenberg’s, Spotify also measures all of their songs with these metrics — Spotify internally records the audio features of all of its songs, as determined by their machine learning algorithm, which can then be accessed through their API.

Using Python and Spotify’s freely available ranking of their top songs, Spotify Charts, I scraped the names of the Top 200 songs on Spotify on each day over the past year — 73,000 songs in total. Then, I used their API to find the arousal (called “energy” by Spotify), and valence of each of those songs. Spotify measures both of these metrics on a 0.0 to 1.0 scale, with 0.0 indicating minimum energy and valence, and 1.0 indicating maximum energy and valence.

After mapping the summary statistics for the energy and valence distributions in Tableau, where red is energy and yellow is valence, we get this — a map of how Americans’ moods have changed over the past year:

Image for post
Image for post
Image for post
Image for post
Image for post
Image for post

The best part is that it’s interactive — you can go to any date you like and see what the musical metrics were like for that day. If you’d like to play around with the interactive version of the visualization, click here. To see the other dates, try the different Sheets at the bottom of the Tableau Public page.

What the Biggest Events of the Year Were, According to Spotify

After calculating outliers and exceptional dates with SPSS and R, we can see when Americans’ musical tastes drastically shifted, and compare that to what was going on in the US at the time, demonstrating a correlation between certain events and personality changes in Americans as a whole. While some of the changes make sense, others were quite surprising:

January 16-January 25: Trump’s impeachment trial begins

Change: (Outlier) Maximum energy hits a high outlier of 0.966.

What This Says About Americans’ Mood:

Energy is positively correlated with neuroticism. Unsurprisingly, as the country waits with bated breath for the results of the trial, their sensitivity to stress increases.

January 27: Kobe Bryant dies

Change: (Notable change) Median energy decreases and median valence increases

What This Says About Americans’ Mood:

Kobe Bryant’s sudden death is met with lower median energy and higher median valence, which are correlated with lower stress and higher openness. Despite the tragedy of the situation, Americans are brought together by their memories of Bryant’s impact on their lives.

January 28-January 31: COVID worsens; WHO declares COVID a “global emergency”

Change: (Outlier) Maximum energy drops down to a low outlier of 0.892

What This Says About Americans’ Mood:

Surprisingly, Americans show decreased neuroticism, or that they’re less sensitive to stress, as the pandemic is widely seen as being especially bad.

Image for post
Image for post

March 13, March 14: Days following WHO declaring COVID a “pandemic”

Change: (Notable change) Energy and valence increase

What This Says About Americans’ Mood:

The combination of increased energy and valence is correlated with not just increased sensitivity to stress, but also to increased conscientiousness, in the wake of this jarring event. Unfortunately, this trend only lasts for about a week, after which the indicators of increased stress and conscientiousness disappear.

March 22- March 23: Senate Fails to Pass COVID Stimulus Bill

Change: (Notable change) Energy and valence decrease

What This Says About Americans’ Mood:

Lower energy and valence are correlated with agreeableness, indicating that Spotify listeners were pleased with the bill’s failure. This might be explained by the bill’s controversial emphasis on sending financial aid to companies over workers.

July 24-July 26: George Floyd protests continue around the world

Change: (Outlier) Maximum energy hits a high outlier of 0.968

What This Says About Americans’ Mood:

Two months after the death of George Floyd, protests not only continue across the country in the US, but have reached a fever pitch after rapidly building momentum. As Americans go protesting during this weekend, they become increasingly stressed at the turmoil of the situation.

Even as 2020 winds down, it will remain a year to remember. Moreover, the music which we listened to will forever memorialize how each of the many significant events in 2020 changed us.

Follow me on Twitter @MellowYellowData

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store