Visualizing Taylor Swift's eras as color swatches
By Danielle Carrick
Taylor Swift famously shifts her sound with each album, creating new "eras" in her discography. From her twangy country origins in Taylor Swift to the indie-infused narratives of folklore and evermore, Swift has continuously reinvented her sound.
Sonically, we recognize these distinct eras in her discography…but can we quantify them? Visualize them? What is it about each album that contributes to its unique sound?
The Spotify API provides a set of audio features for every song that we can use to understand trends through Taylor Swift's discography.
In particular, I use four of Spotify's measures:
- Energy
- Danceability
- Acousticness
- Loudness
To visualize a song, I convert these measures to RGBA space. RGB stands for Red, Green and Blue. A stands for Alpha, which controls opacity. Assigning a measure to each of these channels gives me a color swatch for each song. String all the songs in an album together into a ribbon, and this gives us an impression of the sonic DNA within and across albums.
But what really are these measures? And how exactly do we go from four abstract numbers describing the song to a single color swatch?
Determine a color for each track
More energetic songs appear redder
Energetic songs are characterized by intensity and vigor. Taylor Swift's most energetic song is "Haunted" off of Speak Now. It pulses with emotion and intensity, and presents in the analysis as a vividly warm red.
Haunted (Taylor's Version)
Speak Now (Taylor's Version)
More danceable songs appear green
Danceability measures musical elements like tempo, rhythm stability, beat strength, and overall regularity. "I Think He Knows" off of Lover has one of the highest danceability scores, and and so its corresponding color swatch is a sugary lime green.
I Think He Knows
Lover
More acoustic songs appear blue
Songs with high acousticness will have stripped-down arrangements and organic instrumentation. "Sweet Nothing" off of Midnights is a nice example of a highly acoustic song. Look at its swatch here:
Sweet Nothing
Midnights (The Til Dawn Edition)
I know you might be thinking "Hey, that looks pretty purple". And you're right! That's because while it ranks high on acousticness, it has a low loudness value, making it lighter. Here is what it would look like without opacity applied. Now that's blue!
Things to look for in the Era's Paint Swatchs
As you explore all of the albums below, there are a few things to pay attention to. First of all, you'll notice that folklore and evermore, considered part of the same era, are also visually similar. They both have color swatches composed of light purple with a sprinkle of blues and pinks. Those two albums represent Swift's journey into indie influences and showcase some of her most acoustic songs.
In fact, her last 4 albums have been the most acoustic of her career. Earlier albums have higher median danceability and energy values. The tracks on 1989 (Taylor's Version) and Speak Now (Taylor's Version) both have the highest median energy value resulting in their color ribbon's stronger red tones. Lover has the highest median danceability and therefore the most green tones.
Of course this approach is subject to interpretation. Using different audio features or assigning them to other colors will produce varied color ribbons and highlight unique sounds. I don't think we can truly define a song with four metrics and a color, but even these few features quickly illuminate the musical quality of Swift's work.
Album Color Ribbons
Tap to see each song title.
The Tortured Poets Department
Midnights (The Til Dawn Edition)
evermore (deluxe version)
folklore (deluxe version)
Lover
reputation
1989 (Taylor's Version) [Deluxe]
Red (Taylor's Version)
Speak Now (Taylor's Version)
Fearless (Taylor's Version)
Taylor Swift
Note: To account for multiple releases of an album, the release with with greatest number of tracks is used. In addition, when available, only "(Taylor's Version)" releases are included. No live albums are included. I used a logarithmic scale to convert energy, danceability and acousticness audio features into a value between 0 and 255. The alpha value (also known as opacity) is a number between 0 and 100% and used a linear scale. Support and consultation provided by Stuart Mason.