Tyler, the Creator Text Analysis

EXPLICIT WARNING

Background

The Artist

Since beginning his career over a decade ago as Odd Future’s leader, Tyler, The Creator has forged his own path in hip-hop, crossing genres and surpassing expectations. In 2009, Tyler dropped his debut mixtape, 'Bastard,' referencing vivid scenes of violence and his broken relationship with his father. The LA native’s 2011 debut album, 'Goblin,' generated controversy for its homophobic language, but he followed up with a more well-rounded project with 2013’s 'Wolf' and drew on a variety of musical influences on 2015’s 'Cherry Bomb.' Tyler showed further personal and musical growth on his fourth studio album, 'Flower Boy,' which features the rapper exploring his vulnerability and sexuality. 2019 brought the critically acclaimed No. 1 album, 'IGOR,' which is entirely written, produced, and arranged by Tyler.

Hypothesis

I will analyze the text of all seven of Tyler, The Creator's albums in order to evaluate how his lyrics have evolved over the years. I hypothesize that as his older albums are more negative than his newer albums and that his albums become more positive with as time goes on.

Process

I started my analysis by loading the genius package into R. I then proceeded to grab the lyrics from all of the albums using this package and stored it into its own variable. The code I used for his first album "At Your Own Risk" is shown below. For the sake of brevity, all of the following examples of code will be from this album as well.

atYourOwnRisk <- genius_album(artist="Tyler, the Creator", album = "At Your Own Risk")

I then proceeded to unnest each word from all the songs in the album and filtered out all the stop words. I stored the cleaned out version of the albums into its own variable. Afterwards, I started the sentiment analysis by creating another new variable. (only the code for bing is shown)

ayorWords <- atYourOwnRisk %>% unnest_tokens(word, lyric) ayorCLEAN <- ayorWords %>% anti_join(stop_words) %>% count(word, sort=TRUE) get_sentiments("bing") ayorBING <- ayorCLEAN %>% inner_join(get_sentiments("bing"))

GGPLOTs of Bing Sentiment Analysis

Using the new variables, I created ggplot bar graphs displaying the top negative and positive words used in each variable. This was done after separating the positive and negative words from eachother with the new variable "ayorBING2". The code and the graphs are shown below.

ayorBING2 <- ayorBING %>% group_by(sentiment) view(ayorBING2) ggplot(ayorBING2, aes(word, n, fill= sentiment)) + geom_col(show.legend = FALSE) + facet_wrap(~sentiment, scales = "free_y")+ coord_flip()

At Your Own Risk (2007)

Bastard (2009)

Goblin (2011)

Wolf (2013)

Cherry Bomb (2015)

Flower Boy (2017)

IGOR (2019)

AFINN Sentiment Analysis

I continued the sentiment analysis using AFINN to find a total "score" for each album. This would allow me to see which albums were more positive or more negative and by how much they were. AFINN assigns numbers to each positive word. The more negative the number, the more negative the word and vice versa. I added the scores using the following code.

ayorNEW <- ayorAFINN$n * ayorAFINN$value sum(ayorNEW)

The score for each album:

At Your Own Risk (2007)= 13

Bastard (2009)= -1869

Goblin (2011)= -3403

Wolf (2013)= -2066

Cherry Bomb (2015)= -1012

Flower Boy (2017)= -511

IGOR (2019)= -167

Word Clouds of Every Album

After the sentiment analysis, I created word clouds for each album. This allowed me to see which word(s) were more popular and less popular compared to others in the albums.

wordcloud2(na.omit(ayorCLEAN), size=1, shape='pentagon', fontFamily = "Impact", color = 'random-light', backgroundColor = "#1", rotateRatio = 1)