Sentiment Comparison Between Popular TV Shows
Intro
What makes a TV show popular? While that is a multifaceted question with no straightforward answer, we want to find similarities and differences between three vastly different and very successful TV shows (Gilmore Girls, Euphoria, South Park) to see what makes them so popular. To look into the “why,” we have done text and sentiment analysis of the scripts of these shows and contextualized these analyses with our own interpretations as to why these words and general sentiments might resonate with viewers. Having seen these shows, we are able to confirm whether the analyses accurately depict the essence and topics of the series. First, we had to webscrape these scripts and then we cleaned up the scripts and removed the stop words. We then performed the analyses to reach a conclusion about how and why such disparate TV shows are so appealing to the general public.
Summaries of Each Show
Here are summaries of each of the shows for context before we get into the visualizations and analyses!
Gilmore Girls
Gilmore Girls is a feel-good drama comedy series focused on the relationship between a young single mother and her teenage daughter and their lives in a small New England town in the late 90s/early 2000s. With quirky characters in this town and complex relationships with parents and extended family, Gilmore Girls follows this unique mother-daughter duo through academic, professional, romantic, and existential ups and downs. This show has a very characteristic tone, with a lot of pop culture references and speedy, witty banter in the dialogue. The sense of community, ambition, independence, and love has left a lasting effect on millions of viewers.
Euphoria
Euphoria is a teen drama series known for its graphic depiction of mature topics, especially sexual abuse and drug use, as well as its focus on cinematography and scoring. It follows the lives of Rue and her high-school classmates as they navigate a variety of challenges from self-confidence and sexuality issues to addiction, grief, and abuse. The show’s distinct visual style targets and portrays Gen Z in a way that has significantly influenced trends in fashion, makeup, and social media.
South Park
South Park South Park is a satirical sitcom infamous for its exploration on offensive and controversial topics. The show centers around four fourth-graders (Cartman, Kyle, Stan, Kenny) and their adventures throughout their town of South Park, CO. The story can range from being unwillingly transported to a foreign country to commentaries on American politics.
Webscraping
To obtain these scripts to analyze, we each individually found a website containing the episode scripts and webscraped them into a dataframe.
Gilmore Girls
Gilmore Girls has a total of 157 episodes across 8 seasons, each about 45 minutes long (and the last season is actually just 4 episodes of 1.5 hour length). The scripts for all of these episodes can be accessed on https://scriptmochi.com/tv-series/gilmore-girls, a website with links for each episode. We scraped through all of these episodes, creating a data frame with the link of the episodes, the name of the episode, and the whole script. We manually added the season and then separated the text by character lines and then by word.
Euphoria
Euphoria currently has a total of 18 epsidoes, in 2 seasons of 8 episodes each with 2 special episodes between the two seasons. The scripts for many of the episodes can be found at https://euphoria.fandom.com/wiki/Category:Transcripts_of_Euphoria. However, the website has transcripts only for the first two episodes of the second season, and neither of the special episodes, and we were unable to find another source of transcripts for these episodes that were cleanly and consistently formatted enough to reliably webscrape. Therefore, we focused analysis on the first season and the first two episodes of the second season. We scraped these 10 episodes and created a data frame containing each line and the character who says it by separating by colons, then went back and corrected some irregularities like missing colons and unnecessary line breaks to redo this analysis and fix some incorrectly parsed lines. Then, we separated the lines into their component words, still keeping data on which characters say which words.
South Park
South Park has a total of 321 episodes (excluding specials, movies, etc) that are each around 20-25 minutes long. The Fandom Wikipedia has the scripts for all these shows, so scraping each table from each of these links would get us all the lines. We created a list of links based on the episode titles (scraped from wikipedia) and then iterated through all these links to webscrape all the lines. Most of the episodes were retrieved, but through this method, there were certain episodes we were unable to get (88.1%). However, this is more than enough to get the general idea of South Park, so we continued. Then, we wrangled to get rid of stage directions and split by word.
Text Analysis
First, we wanted to look at the most common words in the show before we use the sentiment analysis to analyze their positivity.
Most Common Words
Here are the most common words in these three shows after removing the stop words (using the dplyr stop words data set and individually selected stop words that were not included in that data set).
Gilmore Girls
Euphoria
South Park
Sentiment Analysis
To implement the sentiment analysis, we used tidy text databases called AFINN and nrc that describe how postive/negative a word is with a score from -5 to 5 and associate a number of categorical sentiments like fear, trust, joy, anticipation, and anger with the same words, respectively. There are only 2,477 words that exist in this database and it cannot take into account context and tone, revealing this analysis’s limits, as well as potentially removing certain words we deemed not a stop word
Most Common Words by Sentiment
Instead of looking just at the most common words, we wanted to filter even more by dplyr’s sentiment package and look at the most common words that have an associated positivity value from the AFINN database. Using a color scale where green is postive and red is negative, we can reevaluate the recurring themes and topics in the shows.
Gilmore Girls
Euphoria
South Park
Positivity Scores
These mean positivity scores represent the overall sentiment of each three shows as a whole. We would expect a neutral show to have an average score of 0, as all the positive (1 to 5) and negative (-5 to -1) scores to cancel out. However, we can see that Gilmore Girls is slightly positive, while South Park and Euphoria are more negative.
| Show | Positivity Score |
|---|---|
| Gilmore Girls | 0.218 |
| Euphoria | -1.373 |
| South Park | -0.519 |
Categorical Sentiment Analysis
Given in the package are these 8 categorical sentiments, where each word in the database is associated to one or more of these sentiments. We joined these sentiment assignments with the words contained in each of these scripts to display the difference in sentiment breakdown for the three different shows.
Gilmore Girls
Euphoria
South Park
Character Analysis
In each of these shows we narrowed down the characters to the top 5 (or 4) main characters that contribute the most to the dialogue or culture of the show. This shows the breakdown of the most common words seen above into the most common characters, highlighting their similarities and differences.
Gilmore Girls
The top 5 characters in Gilmore Girls: Lorelai and Rory, the mother and daughter duo. Rory, a studious and ambitious teenager, takes after her energetic, driven, coffee-obsessed mother Lorelai. Luke, the owner of the town diner that Lorelai and Rory frequent is also Lorelai’s future boyfriend/husband. Emily, Lorelai’s mother and Rory’s grandmother, is a very posh, stubborn woman. Sookie is Lorelai’s bubbly and sometimes frazzled best friend and is the chef at the inn that they work at/run together.
| Character | Filtered Words | Unfiltered Words | % of all Filtered words | % of all Unfiltered words |
|---|---|---|---|---|
| Lorelai | 82981 | 308012 | 26.49 | 26.56 |
| Rory | 46281 | 179903 | 14.78 | 15.51 |
| Luke | 21765 | 87082 | 6.95 | 7.51 |
| Emily | 19464 | 73513 | 6.21 | 6.34 |
| Sookie | 12261 | 45332 | 3.91 | 3.91 |
| Character | Mean Positivity Score Throughout the Show |
|---|---|
| Lorelai | 0.203 |
| Rory | 0.187 |
| Luke | -0.190 |
| Emily | 0.419 |
| Sookie | 0.308 |
Euphoria
The five characters who speak the most in Euphoria are Rue, Nate, Jules, Maddy, and Kat. Rue is a teen struggling with severe drug addiction while exploring her queer identity, and much of the show is seen through her lens. Jules is a transgender girl and Rue’s on-and-off girlfriend. Nate, on the other hand, is a misogynistic star football player with severe anger issues stemming from his sexual insecurity and dysfunctional family. Jules seeks male attention through online dating and hookup apps, through which she inadvertently has romantic and sexual interactions with both Nate and his father, kicking off much of the tension of the first season. Meanwhile, Nate is in a physically and emotionally abusive relationship with Maddy, which complicates many of the consequences of his impulsive and harmful behavior, and Maddy struggles to navigate and escape from his abuse. Lastly, Kat is a girl whose main struggle is her body image, and as she learns to love herself and become more confident, she also turns to illegally working as an online sex worker while still in high school, gaining her money but also validation and self-esteem.
| Character | Filtered Words | Unfiltered Words | % of all Filtered words | % of all Unfiltered words |
|---|---|---|---|---|
| Rue | 3719 | 13846 | 33.52 | 33.84 |
| Nate | 745 | 2833 | 6.71 | 6.92 |
| Jules | 679 | 2813 | 6.12 | 6.87 |
| Maddy | 499 | 1855 | 4.50 | 4.53 |
| Kat | 487 | 1911 | 4.39 | 4.67 |
| Character | Mean Positivity Score Throughout the Show |
|---|---|
| Rue | -1.391 |
| Nate | -2.390 |
| Jules | -0.811 |
| Maddy | -1.679 |
| Kat | -1.558 |
South Park
As mentioned, the four main characters of South Park consists of the four boys (Cartman, Kenny, Stan, Kyle). Cartman is a fat, psychopathic child raised by an enabling single mother. Kenny is usually made fun of for being poor, and is infamous for dying in gruesome ways. Stan is your average American with a nuclear family, although that does not make him normal at all. Kyle is a Jewish-American who conflicts heavily with Cartman over his heritage.
| Character | Filtered Words | Unfiltered Words | % of all Filtered words | % of all Unfiltered words |
|---|---|---|---|---|
| Cartman | 43869 | 140919 | 15.43 | 15.55 |
| Kenny | 1430 | 4461 | 0.50 | 0.49 |
| Kyle | 18986 | 67360 | 6.68 | 7.43 |
| Stan | 19767 | 69018 | 6.95 | 7.61 |
| Character | Mean Positivity Score Throughout the Show |
|---|---|
| Cartman | -0.630 |
| Kenny | -1.058 |
| Kyle | -0.842 |
| Stan | -0.698 |
Analysis and Conclusion
Analysis
Some of the most notable common words in Gilmore girls are time, mom, call, people, love, and coffee, which reveal many of the common themes in the show: the passage of time/growing up, motherhood and mother-daughter relationships, community, love, and of course, coffee. This show is a lot more positive than the other ones by positivity score, the color-coded most common words, and the character breakdown of positivity scores. This makes sense, as it is intended to be a comfort show and does not use too many cuss words or consistently deal with really negative topics.
Regarding the character analysis, it makes sense that Lorelai and Rory say mom a lot since their relationships with their mothers are incredibly formative for them and are crucial to the basis of the show. The lack of the word dad in Rory’s top 15 words also reveals the absence of her father in her upbringing in comparison to Lorelai, where dad is one of her most common words. Emily’s positivity score seems shocking at first since she is well known as a very belligerent and arrogant character at times, but this can be attributed to her mannerisms and the words like wonderful that she tends to use, which signify her high socioeconomic status more than other character traits. Luke has a lower positivity score, which aligns with his grumpy and sarcastic nature. Sookie has a high score as well, as she is very bubbly and goofy.
In Euphoria, all of the main characters and the show as a whole have significantly negative positivity scores, which is unsurprising given its very dark subject matter. Despite this overall negative tone, the story also depicts a wide variety of different types of love, although some of these deep emotional relationships have undesirable consequences. This is evident in the place of the word love as the most commonly used non-swear word in the show, and the swear words can almost be seen as quasi-stop words because of how ubiquitously they are used. It also contains frequent commentary on not only high school life but the modern world in general, as we can see with life, feel, time, god, real, and people also appearing as common words.
Rue, the main protagonist, speaks significantly more than any other character because she also often acts as a narrator. This also makes sense with the fact that her positivity score is the closest to the rating of the series as a whole. Jules is Rue’s on-and-off girlfriend throughout the first season. One of the main plot lines of the season is their internal conflict as they both grapple with how their own identities fit into the relationship, as demonstrated by the appearance of feel in both of their lists of most frequent words. Jules has the highest positivity score of the main characters, as she is often the only source of hope and comfort for Rue during her struggle with addiction. She is also shown texting frequently, which is why ur and lol are among her most used words. Nate’s profoundly troubled and aggressive personality earns him the lowest positivity score with commonly said words like r*pe, hurt, and motherf*cker. Maddy and Kat both speak less than the three main characters, so their vocabularies seem slightly less revealing of their personalities and relationships, but Kat is the only character of the five to have the words sex and cool in her top 15 words, hinting at her issues with body positivity, self-confidence, and her relationship with sex.
For South Park, the overall negative positivity score (both by the show as a whole and by the four main characters) reflects its status as a satirical comedy unafraid to make bold statements. However, that also means that there is a nice balance of heart-warming moments as well (compared to a show like Euphoria which these moments are few and far between). This is reflected on the sentiments based on the most common (non-stop) words and the categorical sentiment analysis (eg. fear and joy were very close in terms of instances). This mix of emotion, as well as the balance of the main characters’ characteristics likely draws and maintains its fanbase.
A top sentiment that all three of these shows had was trust, which makes us curious about what is defined as trust but also highlights that one commonality across all these popular shows is a feeling of trust.
Overall, it is clear that these shows differ a lot based on the most common words, positivity, and sentiment breakdown and reveal the different themes and genres of these shows; these shows cater to a variety of target audiences that seek different viewing experiences and may enjoy different forms of media to pass the time.
Conclusion
Gilmore Girls has been consistently popular since it aired for its comforting, relatable, and almost utopian depiction of a tight-knit small town community. Euphoria is popular for its captivatingly graphic depiction of serious high school problems while also maintaining a consistent visual aesthetic through its lighting, makeup, and fashion. People like South Park for its unapologetic tellings of modern stories with a balanced character selection that isn’t afraid to say what they want to say, but aren’t complete potty mouths. These three different shows reveal that humans tend to consume media for different purposes: comfort, escape, humor, shock, and drama, to name a few. They show different aspects of human nature and the way that modern media is consumed for numerous purposes.
For future analyses, using some sort of machine learning to get a sentiment analysis tailored to shows rather than being a one-size-fits-all database would be more effective at showing context behind certain words. Although we agree with most of the values associated to words in the AFINN database, words such as b*tch were placed more negative than words like f*ck, which does not account for these words being used in a positive way. Also, the AFINN database is very limited, as it only has 2,477 words. A more extensive sentiment analysis with more words and a more accurate analysis of their positivity could shed more light on the sentiments in the show.
Overall, this deep dive into this selection of popular media was demonstrative of humans’ desire to watch other humans interact in a plethora of different flavors and ways that they may not experience in their daily lives.