Taylor Swift Discography Exploration

Author

Paige Gunther

Introduction

My name is Paige Gunther and I am a senior Business Analytics and Marketing double major. One of my main hobbies is listening to music. As long as I can remember, I have found connection through music and artists lyrics. It amazes me how one individual can write about something so personal and have millions of people connect to it in their own way. One artist I find does this exceptionally well is Taylor Swift. I have been listening to her for as long as I can remember!

For my data exploration in my final capstone for Business Analytics, I wanted to explore Swift’s discography. With such an expansive discography, I was sure there are a plethora of things to uncover. The data set I chose mostly stems from Spotify data, looking at each song of Taylor Swift’s and identifying different variables listed below:

  • album name

  • album genre

  • album announcement date

  • album release date

  • track number

  • track name

  • artist

  • featuring artists

  • bonus track (True or False if it was a bonus track on the album)

  • track release date

  • danceability score (0-1) (how suitable is the track for dancing)

  • energy score (0-1) (Represents a perceptual measure of intensity and activity)

  • loudness score

  • speechiness score (0-1) (measures the presence of spoken words in a track)

  • accounticness score (0-1) (measures how much of a song’s sound is produced using non-electronic, acoustic instruments)

  • liveness score (0-1) (Detects the presence of a live audience in the recording)

  • valence score (0-1) (measuring the “positiveness” of a track)

  • tempo

  • time signature

  • duration of the song (in ms)

  • explicit (True or False if the song contains explicit language)

  • key name

  • mode name (major or minor)

Descriptive Analysis

I created a set of visuals to understand my data better and draw conclusions about Swift’s discography.

The first thing I looked at was the different tempos of songs in each genre Taylor Swift has in her discography. Based on the box plot below, we can see that many of the median tempos for each genre appear to be around the same. However, country-pop tends to have more condensed, similar tempos across songs while electropop has many different tempos in the genre!

I then went on to look at which track number is, on average, the longest. From this bar graph, we can see that the 30th track number is, on average, the longest. However, Taylor only has two albums with a 30th track number. They often have closer to 18 tracks. The 30th track on her album Red is a 10-minute version of another song, accounting for the outlier. We can see that, if we ignore that track, the 22nd track is longest. Even still, we see that Taylor writes longer songs than the average artist. Fun fact: the average song length is 3 to 3.5 minutes.

Are shorter songs more “danceable” than longer songs?

This scatterplot shows that there is no real relationship between song length and song danceability. I omitted the 10-minute version song because it acts as an outlier.

The dancability variable was incredibly interesting to me so I dug in further. The next question I asked was: What is the relationship between tempo and dancability? Since there appears to be no clear relationship between song length and dancability, I was curious what relationship dancability has with the rest of the data. Tempo makes sense and is something I wanted to confirm, but the graph I created surprised me. According to this visual, I see that as a song’s tempo increases, the dancability actually decreases. This means Taylor’s faster songs are not necessarily her more dancable songs. This could mean that she has slower songs that are good for slow dance styles or that more up-tempo songs have less consistent rhythm.

One common perception people have of Taylor Swift’s music is that she only writes “break up” songs. This gives her songs a negative connotation, so I wanted to use data to either back up or refute this claim. First, I found the average valence of her songs to see if she uses more cheerful, positive words, or negative, depressing lyrics in her songs. Based on this data set, I found that the average valence is 0.39, supporting the claim that Swift writes less cheerful lyrics.

Based on this outcome, I wanted to look at her valence across albums to see if different stages of her life and career produced different song outcomes. Based on this, we can see that the valence of songs actually varies quite a bit across albums! Midnights has, by far, the lowest average valence, while Lover has the highest average valence. However, this graph also shows us that each album has a wide variety of emotional exploration in it, especially her album 1989!

Secondary Data

I also explored ethical web scraping in this course, so the second portion of my data is a collection of the awards Taylor Swift has won from her discography over the years. The variables explored in this data are:

  • Award

  • Year

  • Recipient (Swift or Album)

  • Category

  • Result (Nominated or Won)

  • Album (Album nominated)

  • Song (Song nominated)

  • Win (binary if she won or not)

There are numerous things I could explore with this data set. One thing I was eager to look at was which type of award Taylor has won the most. However, there are hundreds of unique awards that Taylor has been nominated for, so instead I just looked at awards she has been nominated for 30 or more times. This was a much more manageable number of 11 awards. From the graph below, we can see that Taylor Swift has been nominated the most for Billboard Music Awards with over 100 unique nominations!

I then pivoted to looking at how often Taylor won compared to when she was nominated. I found that, on average, Taylor Swift wins 46.79% of the awards she is nominated for. I then wanted to look at which year she won the most awards in to compare to her discography and see when she had the most “Winning” Album. Here, I see that 2015 appears to be her most nominated year as this is both the most Nominations she did not win, as well as the second highest year for awards she did win. The previous October (October 2014), Taylor released her album 1989. This would indicated that this is one of her most popular albums. The year that Taylor won the most albums overall, however, was in 2024. This was the year she released The Tortured Poets Department, her longest album to date. This was the year Swift was on her largest, record-breaking tour, The Eras Tour.

      
        0  1
  2007  2  4
  2008  5  9
  2009 14 28
  2010 34 29
  2011 29 31
  2012 38 24
  2013 62 43
  2014 33 16
  2015 72 60
  2016 37 29
  2017 11 13
  2018 46 20
  2019 40 35
  2020 43 26
  2021 34 25
  2022 28 31
  2023 51 54
  2024 45 65
  2025 26 28
  2026 14 14

Closing Takeaways

With just these two data sets it is clear that there are many things to be discovered when it comes to Taylor Swift, her discography, and her accomplishments. The visuals I was able to create pertaining to her discography told me a lot about Swift’s range as an artist.

  • Taylor Swift is able to write across 6 distinct genres.

  • Taylor Swift has written many songs longer than the average song length.

  • Taylor Swift’s song length has no correlation to the danceability to the song, but her slower songs tend to be more danceable.

  • Taylor Swift, on average, writes more depressing, less hopeful lyrics for her albums (all average valences are below 0.5) but does provide range in her album valence from song to song.

The awards data set was just as interesting in proving how successful Taylor is as an artist.

  • She has been nominated for hundreds of distinct awards but has been nominated for over 100 Billboard Music Awards

  • The most successful years of Taylor Swift’s career in terms of winning awards have been 2015 and 2024