Gender representation in animation, movies, and TV shows has long been a central concern among critics of popular culture. This story takes a data-driven approach to the conversation by analyzing the most beloved characters on MyAnimeList. Using web scraping, we collected detailed information on every character who received over 1,000 votes, offering a closer look at who fans favor—and what that might reveal about gender dynamics in anime fandom..
The dataset includes:
- Name of the character
- Age
- Occupation
- Height
- etc
First, we'll look at height.
Resutls show that height is not the determing force of favoritism. The most welcomed character Lelouch is around 178 cm.
A more extreme example is Levi, who is only 160cm, an unconventional height for a male character that wins high popularity.
Meanwhile, characters who win more votes are mostly from animetions that have longer series and more episodes, like Code Geass from 2006, and One Piece from 1997.
Interestingly, male characters displayed a greater range of height distributions while the range for female characters tended to be narrower.
Not only that there is few very tall characters, but there is also fewer characters with very short height compared to males.
Second, we will look at the age.
The dataset only extracted people within 100 years old, excluding outliers because it aimed to compare character designs of real human beings.
Female characters mostly fall within the age group of 10-20 years old, and there are relatively fewer characters in their 30s.
By contrast, even though male characters have a similar median age, they tend to have more appearances in middle age groups.
The narrow age range in female character representation often results in the portrayal of young teenage girls who conform to highly specific beauty standards—typically featuring pale skin, large eyes, and overtly feminine traits.
These recurring characteristics contribute to a stereotypical and limited depiction of the "anime girl" archetype.
Anime female characters have long been criticized for its lack of diversity.
This research highlights that character development in animation continues to adhere to narrower standards and demonstrates less creative depth when it comes to female protagonists.
Critics argue that this reflects systemic gender bias within the industry, as well as a broader artistic limitation that hinders the creation of complex and fully realized female characters.
It is worth noticing that among mostly conventional pretty girls with pronunced feminine features, young audience are embracing more diverse chracter design more.
For example, Mikasa Ackerman from Attack on Titan is a tall, muscular girl with a distinctively unconventional feminine appearance, and she is a popular character among young audiences.
Methodology
This project involved a scraping-intensive approach to analyze user preferences for anime and manga characters on MyAnimeList. Only characters who received more than 1,000 votes were included in the dataset to ensure data quality and relevance.
To begin, I used Beautiful Soup to scrape essential details about each character from the main character listing page. This included the character’s name, a link to their profile, the number of votes they received, and the anime or manga title they are associated with.
I then employed Playwright to automate the process of visiting each character's individual page. This allowed me to extract additional information such as the character's full biographical description and the source link for their profile image.
The collected data was compiled and saved into a structured .csv file, which served as the foundation for further analysis.
Since gender was not explicitly stated on the MyAnimeList character pages, I developed a method to infer gender based on the character's biographical text. Using regular expressions, I extracted attributes such as age, birthday, nationality, height, occupation, and a short paragraph describing the character.
To infer gender, I created a keyword-based scoring system. This system scanned the biographical text for female-associated words (such as she, her, woman, girl, female, and lady) and male-associated words (such as he, him, man, boy, male, and gentleman). Each time a keyword was detected, a score was added to the corresponding gender category.
The character's gender was then inferred based on the higher of the two scores. This inferred gender was added as a new column titled gender in the final dataset.
