Negative Sentiment of Slipknot & Stone Sour

Hypothesis

Corey Taylor is the lead singer and song writer for two popular American rock bands: Slipknot and Stone Sour. Both bands are known for their deep lyrics that usually carry negative connotations. However, a portion of listeners believe that the music for Slipknot is more negative for the shock value of being a heavy metal band while the lyrics of Stone Sour songs are more mellow. I hypothesize the Corey Taylor has written similar lyrics for both bands, and that both bands carry similar negative connotations in the lyrics. I also hypothesize that the lyrics for both bands have shown to be more negative over the years.

Method and Analysis

Required R Packages

library(genius)
library(wordcloud2)
library(tidytext)
library(tidyverse)
library(ggthemes)
library(devtools)

To begin the analysis, I pulled the lyrics from 6 chosen albums using the Genius R package. These albums were chosen due to the fact that each band released the albums at a similar time, and that each album contains newly written music instead of previously released songs or covers of other artists.

stoneSourAlbums <-tribble(
  ~ artist, ~ title,
  "Stone Sour", "Come What(ever) May",
  "Stone Sour", "Audio Secrecy",
  "Stone Sour", "Hydrograd",
)

stoneSourLyrics <- stoneSourAlbums %>% 
  add_genius(artist, title, type = "album")

slipknotAlbums <- tribble(
  ~ artist, ~ title,
  "Slipknot", "Vol. 3: (The Subliminal Verses)",
  "Slipknot", "All Hope Is Gone",
  "Slipknot", "We Are Not Your Kind",
)

slipknotLyrics <- slipknotAlbums %>% 
  add_genius(artist, title, type = "album")

I then did 3 separate analysis on each pair of albums. Please see the code for each album in the analysis section. For each pair:

1. I analyzed the top 10 words used in both albums with a bar chart and a word cloud.

2. I used the AFINN sentiment lexicon to identify the negativite sentiment of the lyrics. Not all of the words from the top 10 reoccurring words had a sentiment value in the AFINN Lexicon, so I chose to analyze the top 10 reoccurring words that do have included values in the lexicon.

3. I added the sentiment values of the top 10 reoccurring words that have sentiment values in the AFINN Lexicon to assign each album a sentiment score. I then compared them to the other band’s album released in the similar time period (2-3 years apart).

These are the albums that I analyzed and compared:

Come What(ever) May (2006) by Stone Sour versus Vol. 3: (The Subliminal Verses) (2004) by Slipknot

Audio Secrecy (2010) by Stone Sour versus All Hope is Gone (2008) by Slipknot

Hydrograd (2017) by Stone Sour versus We are Not Your Kind (2019) by Slipknot

Analysis 1

---

Come What(ever) May - Stone Sour, 2006

stoneSourLyrics %>% 
  filter(title %in% "Come What(ever) May") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>%
  filter(!word %in% "30") %>%
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  wordcloud2(size=.9, color='random-light', backgroundColor="black")

stoneSourLyrics %>% 
  filter(title %in% "Come What(ever) May") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "30") %>%
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  head(10) %>% 
  ggplot(aes(reorder(word, -n), n)) + geom_col(fill="#0b7c76") -> tenComeWhatever

tenComeWhatever + 
  ggtitle("Come What(ever) May: Top 10 Words") + theme(plot.title = element_text(hjust = 0.5)) + scale_color_fivethirtyeight() + theme_fivethirtyeight() 

stoneSourLyrics %>% 
  filter(title %in% "Come What(ever) May") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "30") %>%
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  inner_join(get_sentiments("afinn")) -> comeSent

comeSent %>% 
  head(10) 

## # A tibble: 10 x 3
##    word       n value
##    <chr>  <int> <dbl>
##  1 die       14    -3
##  2 god       11     1
##  3 care       9     2
##  4 stop       8    -1
##  5 afraid     7    -2
##  6 free       7     1
##  7 refuse     6    -2
##  8 shit       6    -4
##  9 tired      6    -2
## 10 fight      5    -1

For my analysis of Stone Sour’s 2006 album, Come What(ever) May, I chose to not include the top reoccurring words “30” and “yeah”. The word “yeah” is primarily used as a filler word in the songs. The number “30” is referring to the lead singer’s gene size and weight as way to identify himself in a song about his high school years. For example: “30/30-150 Remembers 30/30-150 HATES.” Many of the top used words do not have sentiment assignments in the AFINN Lexicon, so I instead calculated the sentiment score using the top reoccurring words with assigned sentiment values.

Vol. 3: (The Subliminal Verses) - Slipknot, 2004

slipknotLyrics %>% 
  filter(title %in% "Vol. 3: (The Subliminal Verses)") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  wordcloud2(size=1.2, color='random-light',backgroundColor="black")

slipknotLyrics %>% 
  filter(title %in% "Vol. 3: (The Subliminal Verses)") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  head(10) %>% 
  ggplot(aes(reorder(word, -n), n)) + geom_col(fill="#e69e5b") -> tenVolThree

tenVolThree + 
  ggtitle("Vol. 3: (The Subliminal Verses): Top 10 Words") + theme(plot.title = element_text(hjust = 0.5)) + scale_color_fivethirtyeight() + theme_fivethirtyeight() 

slipknotLyrics %>% 
  filter(title %in% "Vol. 3: (The Subliminal Verses)") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  inner_join(get_sentiments("afinn")) -> volThreeSent 

volThreeSent %>% 
  head(10) 

## # A tibble: 10 x 3
##    word       n value
##    <chr>  <int> <dbl>
##  1 die       15    -3
##  2 forget    15    -1
##  3 leave     12    -1
##  4 insane    11    -2
##  5 fight      8    -1
##  6 save       8     2
##  7 love       6     3
##  8 pain       6    -2
##  9 stop       6    -1
## 10 alive      5     1

For my analysis of Slipknot’s 2004 album, Vol. 3: (The Subliminal Verses), I chose to not include the word “yeah”. The word “yeah” is primarily used as a filler word in the songs. Similar, to the Stone Sour Album in this analysis, many of the top used words do not have sentiment assignments in the AFINN Lexicon, so I instead calculated the sentiment score using the top reoccurring words with assigned sentiment values.

Analysis 1 Results:

Stone Sour sentiment score total for top reoccurring words: -4

Slipknot sentiment score total for top reoccurring words: -5

The sentiment scores reveal that the albums by both bands rank in the negatives for the sentiment value when using the AFINN Lexicon. There is a small difference between the two albums, with Slipknot scoring a stronger negative by one point. The word “die”, which has a negative sentiment score, is in the top 10 reoccurring words for both albums. Both albums also share the words “life” and “feel”/“feels”, but these words to not have an assigned value in the AFINN Lexicon. Overall, this analysis indicates that the sentiment of both albums are very similar in regards to negativity.

Analysis 2

---

Audio Secrecy - Stone Sour, 2010

stoneSourLyrics %>% 
  filter(title %in% "Audio Secrecy") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = TRUE) %>% 
  wordcloud2(size=.9, color='random-light', backgroundColor="black")

stoneSourLyrics %>% 
  filter(title %in% "Audio Secrecy") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = TRUE) %>% 
  head(10) %>% 
  ggplot(aes(reorder(word, -n), n)) + geom_col(fill="#0b7c76") -> tenAudio

tenAudio + 
  ggtitle("Audio Secrecy: Top 10 Words") + theme(plot.title = element_text(hjust = 0.5))+ scale_color_fivethirtyeight() + theme_fivethirtyeight() 

stoneSourLyrics %>% 
  filter(title %in% "Audio Secrecy") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = TRUE) %>% 
  inner_join(get_sentiments("afinn"))  -> audioSent

audioSent %>% 
  head(10) 

## # A tibble: 10 x 3
##    word        n value
##    <chr>   <int> <dbl>
##  1 die        16    -3
##  2 save       14     2
##  3 wrong      11    -2
##  4 hate       10    -3
##  5 haunt      10    -1
##  6 destroy     9    -3
##  7 love        7     3
##  8 true        7     2
##  9 free        6     1
## 10 honest      6     2

For my analysis Stone Sour’s 2010 album, Audio Secrecy, there were no reoccurring words that needed to be removed from the analysis. However, many of the top ten reoccurring words did not have values assigned in the AFINN Lexicon, so I instead used the top 10 reoccurring words that did have values to calculate the sentiment score.

All Hope Is Gone - Slipknot, 2010

slipknotLyrics %>% 
  filter(title %in% "All Hope Is Gone") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "hey") %>%
  count(word, sort = TRUE) %>% 
  wordcloud2(size=1.2, color='random-light',backgroundColor="black")

slipknotLyrics %>% 
  filter(title %in% "All Hope Is Gone") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "hey") %>%
  count(word, sort = TRUE) %>% 
  head(10) %>% 
  ggplot(aes(reorder(word, -n), n)) + geom_col(fill="#e69e5b") -> tenAllHope

tenAllHope + 
  ggtitle("All Hope Is Gone: Top 10 Words") + theme(plot.title = element_text(hjust = 0.5)) + scale_color_fivethirtyeight() + theme_fivethirtyeight()

slipknotLyrics %>% 
  filter(title %in% "All Hope Is Gone") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "hey") %>%
  count(word, sort = TRUE) %>% 
  inner_join(get_sentiments("afinn")) -> allHopeSent

allHopeSent %>% 
  head(10)

## # A tibble: 10 x 3
##    word        n value
##    <chr>   <int> <dbl>
##  1 pretend    24    -1
##  2 dead       21    -3
##  3 die        21    -3
##  4 save       16     2
##  5 hope       15     2
##  6 free       13     1
##  7 care       12     2
##  8 war        10    -2
##  9 god         8     1
## 10 killing     8    -3

For the analysis of Slipknot’s 2008 album, All Hope Is Gone, I removed the word “hey” from the analysis because it is used as a filler word in several of the songs. There continues to be some words in the top ten that do not have sentiment values assigned to them, so I again just used the top 10 words that do have sentiment values to analyze the total sentiment score.

Analysis 2 Results:

Stone Sour sentiment score total for top reoccurring words: -2

Slipknot sentiment score total for top reoccurring words: -4

The sentiment scores again reveal that the albums by both bands rank in the negatives for the sentiment value when using the AFINN Lexicon. There is still small difference between the two albums, with Slipknot scoring a stronger negative by two points. The word “die”, which has a negative sentiment value, continues to be in the top 10 reoccurring words for both albums. The word “free” is one of the top ten reoccurring words with a positive sentiment value in both albums. Overall, this again indicates that the sentiment of both albums are very similar in regards to negativity. It is notable that Slipknot’s album now ranks 2 points more negatively than Stone Sour’s album. However, the sentiment in both albums is less negative compared to the earlier released albums from the first analysis.

Analysis 3

---

Hydograd - Stone Sour, 2017

stoneSourLyrics %>% 
  filter(title %in% "Hydrograd") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  wordcloud2(size=1.7, color='random-light', backgroundColor="black")

stoneSourLyrics %>% 
  filter(title %in% "Hydrograd") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  head(10) %>% 
  ggplot(aes(reorder(word, -n), n)) + geom_col(fill="#0b7c76") -> tenHydrograd

tenHydrograd + 
  ggtitle("Hydrograd: Top 10 Words") + theme(plot.title = element_text(hjust = 0.5)) + theme_dark()  + scale_color_fivethirtyeight() + theme_fivethirtyeight() 

stoneSourLyrics %>% 
  filter(title %in% "Hydrograd") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  filter(!word %in% "yeah") %>%
  count(word, sort = TRUE) %>% 
  inner_join(get_sentiments("afinn")) -> hydrogradSent

hydrogradSent %>% 
  head(10) 

## # A tibble: 10 x 3
##    word             n value
##    <chr>        <int> <dbl>
##  1 love            14     3
##  2 feeling         13     1
##  3 dead            12    -3
##  4 shit            11    -4
##  5 fucking         10    -4
##  6 bad              9    -3
##  7 motherfucker     8    -5
##  8 saved            8     2
##  9 fire             7    -2
## 10 fuck             7    -4

For my analysis of Stone Sour’s 2017 album, Hydrograd, I chose to not include the word “yeah”. The word “yeah” is primarily used as a filler word in the songs. Again, many of the top used words do not have sentiment assignments in the AFINN Lexicon, so I instead calculated the sentiment score using the top reoccurring words with assigned sentiment values.

We Are Not Your Kind - Slipknot, 2019

slipknotLyrics %>% 
  filter(title %in% "We Are Not Your Kind") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = TRUE) %>% 
  wordcloud2(size=1.2, color='random-light',backgroundColor="black")

slipknotLyrics %>% 
  filter(title %in% "We Are Not Your Kind") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = TRUE) %>% 
  head(10) %>% 
  ggplot(aes(reorder(word, -n), n)) + geom_col(fill="#e69e5b")-> tenKind

tenKind + 
ggtitle("We Are Not Your Kind: Top 10 Words") + theme(plot.title = element_text(hjust = 0.5))+ scale_color_fivethirtyeight() + theme_fivethirtyeight() 

slipknotLyrics %>% 
  filter(title %in% "We Are Not Your Kind") %>% 
  unnest_tokens(word, lyric) %>% 
  anti_join(stop_words) %>% 
  count(word, sort = TRUE) %>% 
  inner_join(get_sentiments("afinn")) -> kindSent

kindSent %>% 
  head(10) 

## # A tibble: 10 x 3
##    word      n value
##    <chr> <int> <dbl>
##  1 love     25     3
##  2 death    18    -2
##  3 liar     17    -3
##  4 kill     10    -3
##  5 pain      9    -2
##  6 worst     9    -3
##  7 hell      8    -4
##  8 hope      8     2
##  9 enemy     7    -2
## 10 fuck      7    -4

For my analysis Slipknot’s 2019 album, We Are Not Your Kind, there were no reoccurring words that need to be removed from the analysis. However, like all of the analysis above, many of the top ten reoccurring words did not have values assigned in the AFINN Lexicon, so I again instead used the top 10 reoccurring words that did have values to calculate the sentiment score. However, it is important to note that the top three reoccurring words (“love”, “death”, and “liar”) all do have values on the AFINN lexicon, with “love” being the only word that scores positively.

Analysis 3 Results:

Stone Sour sentiment score total for top reoccurring words: -19

Slipknot sentiment score total for top reoccurring words: -18

The sentiment scores again reveal that the albums by both bands rank in the negatives for the sentiment value when using the AFINN Lexicon. There is still small difference between the two albums, with Stone Sour scoring a stronger negative by one point. This is notable because Slipknot scored more negatively in the prior analysis. The word “love”, which has a positive sentiment value, is in the top 10 reoccurring words for both albums. The words “death”/“dead” and “fuck”/“fucking” are in the top ten reoccurring words with an assigned sentiment value(negative, in this case) in both albums. Overall, this again indicates that the sentiment of both albums are very similar in regards to negativity. However, the sentiments of both albums are far more negative compared to the earlier released albums from the first and second analysis, with over a 15 point decrease in sentiment value. This indicates that the albums released in 2017 and 2019 are far more negative in sentiment, but continue to rank similarly in negativity.

Conclusion

---

By looking at the most common words and sentiment scores for each album in all three analysis, a conclusion can be drawn that Corey Taylor, lead singer and song writer for both Slipknot and Stone Sour, writes similar lyrics with similar negative sentiments for both bands, which supports part of my hypothesis. Some of the most common words are shared between the albums, and the sentiment scores for each analysis were no more than 2 points apart. It is possible that people believe Slipknot’s music is significantly more negative because it is classified as heavy metal while Stone Sour is classified as hard rock. The difference in musical style may also contribute to this belief because the style and sound of Slipknot can be described as very “aggressive” compared to Stone Sour. However, the data shows that the sentiment values for the top used words in each album released within 2-3 years of each other do not differ significantly.

However, the portion of my hypothesis stating the sentiments of both bands have been more negative over the years is not fully supported due to the dip in negativity in albums released in 2008 and 2010 (Audio Secrecy and All Hope is Gone). Stone Sour’s sentiment score increase by two points and Slipknot’s score increased by one point, which means that both albums were less negative in connotation.

The most interesting trend found in this data is the large increase in negativity, as shown by a decreased sentiment score, for both bands in the years 2017 and 2019. Both scores decreased by more than 15 points compared to the prior analyzed albums. This indicates that both bands did become more negative over time, but not as I suspected in my hypothesis. However, it is notable that even with the rise in negativity, both bands only had one-point difference in their negativity score. This could be due to Corey Taylor’s current style of song-writing or his own beliefs and frustrations. He has expressed that he has struggled with his mental health. He has also been known to speak negatively against the current presidential administration, which started in 2016. However, there is no data that ties these trends to any outside factors.