Lyrical Visualization and Analysis: Taylor Swift

Cover Image

I am a longtime fan of Taylor Swift, and I admire her songwriting skills and ability to tell vivid stories with so few words. With a current discography of nine unique albums (not including rereleases or special editions) chronicling her life from ages 15 to 30, Swift has written an extensive amount of song lyrics that are ripe for analysis.

For my first foray into Python programming, I endeavored to take on this vast dataset of song lyrics. I defined my dataset to include Swift’s main studio albums, as follows: Taylor Swift, Fearless, Speak Now, Red, 1989, reputation, Lover, folklore, evermore.

Goals:

  1. Visually model the lyrics, sorted by album, in word clouds (also called tag clouds), which are an eye-catching and effective way to visualize language data
  2. Perform analysis on the dataset to determine lyrical and thematic consistencies and changes throughout Swift's career, and implications of her as a songwriter and wordsmith
  3. Perform rule-based sentiment analysis on the dataset to determine the sentiment or mood of each song and album

  1. Get data: get_songs.pdf
  2. Create word clouds: make_wordclouds.pdf
  3. Perform analysis: sentiment_analysis.pdf
Wordcloud for Album: Taylor Swift
Album: Taylor Swift
Wordcloud for Album: Fearless
Album: Fearless
Wordcloud for Album: Speak Now
Album: Speak Now
Wordcloud for Album: Red
Album: Red
Wordcloud for Album: 1989
Album: 1989
Wordcloud for Album: reputation
Album: reputation
Wordcloud for Album: Lover
Album: Lover
Wordcloud for Album: folklore
Album: folklore
Wordcloud for Album: evermore
Album: evermore

Type word in all lowercase and press enter.

I identified lyrics in the word clouds that are unique to individual albums, and determined album themes using these words.

  • Taylor Swift
  • 'amen', 'good heart'
    Innocence, belief in God
    'anyone', 'forgiveness'
    Desperation for others
    'can't', 'weakness'
    Lack of ability
    'tear apart', 'undone'
    Ruined and broken
    'black dress', 'chest', 'date', 'Drew', 'slow'
    Blooming romance
    'lake', 'moon', 'outside', 'sky'
    Nature
    'Georgia', 'old faded blue jeans', 'porch', 'shotgun', 'slamming screen door'
    Small town Americana
    'guitar', 'radio', 'tapping', 'Tim McGraw'
    Music, specifically country music
  • Fearless
  • 'about', 'fairytale', 'princess'/'prince', 'Romeo'/'Juliet', 'story', 'white horse'
    Romantic fantasy and fiction
    'deep breath', 'fearless', 'head first'
    Inner strength and bravery
    'help', 'scared'
    Struggle and fear
    'storm', 'temper'
    Chaos and volatility
    'angel', 'hallelujah'
    Heavenly God
    'along', 'clean break', 'drag', 'tired'
    Burn out, unwillingness to continue
  • Speak Now
  • 'church', 'crowd', 'marrying', 'vow', 'white veil occasion'
    Wedding
    'enchanted'/'enchanting', 'fate'
    Mystical and spiritual
    'careless'/'careful', 'innocent', 'messed', 'rudely barge in', 'wild'
    Trouble maker versus rule follower
    'figured', 'imagined', 'someday', 'whenever'
    Expectations, the hypothetical future
    'daughter', 'family'
    Familial significance
  • Red
  • 'dressed to the nines', 'marvelous tune'
    Old timey, vintage, bygone
    'confused', 'realizing'
    Learning hard truths
    'impossible', 'slope', 'treacherous', 'trouble'
    Insurmountable challenges, instability
    'Wednesday', 'yesterday'
    Recalling and reminiscing
    'new Maserati', 'top'
    Wealth and status
  • 1989
  • 'clean', 'clear', 'done', 'solve'
    Overcoming or being free of something
    'bad blood', 'flames', 'torture', 'warn'
    Malice and toxicity
    'insane', 'reckless'
    Craziness
    'blank space', 'ex-lovers', 'filled'
    A romantic void
    'hide', 'lock', 'trace'
    Evading detection, staying out of sight while being pursued
    'New York', 'nice dress', 'style'
    Fashion, aesthetic quality
    'grab', 'palm'
    Holding hands, exploration
    'heaven', 'incredible', 'wildest dreams'
    Untouchable fantasies
  • reputation
  • 'bad feeling', 'blame', 'mistakes'
    Ominous anticipation, negative consequences
    'actress starring', 'drama', 'reputation'
    The dark façade of fame
    'bar', 'drinking', 'drug'
    Letting go of inhibitions
    'bought', 'gorgeous', 'nice things'
    Money and superficial materialism
    'chill', 'patience', 'silence'
    Calm and quiet
    'end game', 'precede':
    Beginning/end, before/after
    'everyone', 'them'
    Other people, herd mentality
    'death trap', 'getaway car', 'tied'
    Escaping from captivity
    'check', 'trust/truth'
    Trust and mistrust
    'king/kingdom', 'Lord'
    A higher ruling power
  • Lover
  • 'bless', 'blind faith', 'false God', 'worship'
    Surrendering to belief in something greater
    'combat', 'loud'
    Noisy conflict
    'calm', 'daylight', 'skipping down'
    Feeling content, positive and giddy
    '16th Avenue', 'Cornelia Street', 'London'
    The impact of places on experiences
  • folklore
  • 'angry', 'cursing', 'fuck'
    Expressing aggravation
    'broken', 'canceled', 'exile', 'none', 'ruining everything', 'slipped away'
    Losing it all, self-sabotage
    'highest heels', 'old cardigan'
    The imagery of clothing
    'marvelous time', 'mirrorball', 'party', 'spinning'
    Dancing at a soiree
    'assume', 'seem'
    Unfounded judgments
    'August', 'garden', 'pool'
    Backyard in late summertime
  • evermore
  • 'celebrated', 'champagne', 'glass'
    Drinking to commemorate
    'alive', 'bent', 'bone', 'ivy', 'roots', 'tree', 'wherever', 'willow'
    Organically growing and entwining
    'ship', 'tires', 'train'
    Transportation
    'cost', 'pay'
    Compensation
    'flush', 'water'
    Rinsing and cleansing
    'hollow', 'less', 'stray', 'wreck'
    Being left with nothing
    'across', 'leads', 'open'
    Access to new opportunities
    'motion', 'pause'
    Stop and go
    'prove', 'sure'
    Establishing certainty
    'gold rush', 'stone'
    Earth’s treasures
    'somehow', 'tolerate'
    Reluctant endurance

I identified lyrics in the word clouds that are common among all albums.

  • 'around'
  • 'baby'
  • 'back'
  • 'cause'
  • 'come'
  • 'didn't'/'don't'
  • 'go'/'gonna'
  • 'good'
  • 'got'
  • 'know'
  • 'look'
  • 'love'
  • 'made'/'make'
  • 'never'
  • 'night'
  • 'now'
  • 'one'
  • 'said'/'say'
  • 'see'/'seen'
  • 'still'
  • 'take'
  • 'thing'
  • 'think'
  • 'time'
  • 'wanna'/'want'

I consolidated the themes of each album into thirteen universal themes. The reoccuring themes are present on all albums, while the recent themes are present only on later albums.

  • Reoccuring Themes
  • Brokenness
    Chaos and instability
    God
    Music and dancing
    Nature
    Romance
    Struggle and overcoming
  • Recent Themes
  • 21+
    Fame
    Fashion
    Memories
    Money
    Trust

I analyzed the sentiment or mood of each song and album using VADER, a natural language sentiment analysis tool. VADER labels the sentiments of words as either negative, neutral, or positive, and calculates the proportion of the text that is negative vs neutral vs positive. VADER also calculates a compound score of the sentiment of the text on a scale of -1.0 (most negative) to 1.0 (most positive). I determined the negative, neutral, positive, and compound scores for each song and album.

Dataframe for Songs
Abbreviated table of songs and scores

Compound Sentiment Counts

Maximum and Minimum

Average Scores

Dataframe for Albums
Table of albums and scores
Plot for Albums
Plot of album scores