Methods
To investigate these questions I did a combination of data analysis and interacting in fan spaces. For the data analysis, I collected 2,500 Supergirl fanfiction pieces from AO3. To do this I first went to the “Supergirl (TV 2015)” fandom and used the filter bar to eliminate any non-English works along with works that were incomplete. I then sorted these works by kudos, which is what I deemed the best way to determine popularity of a work. Popularity means that something is liked or admired by many people.
The reason I deemed kudos a better measure of popularity than hits, comments, and bookmarks is it guarantees that a person has read it. Hits means that a person clicked on the story, it doesn’t guarantee that a person has read it nor liked it. Bookmarks usually mean one of two things, one being that they read the work, liked it, and are saving it so they don’t lose it the second being they haven’t read it and want to come back to it. This makes bookmarking a bad measure of popularity because there is no guarantee that the person liked the work. While comments usually guarantee a person read it, it does not mean a person liked the work. People leave comments on works they don’t enjoy all the time. Comment numbers are also inflated by the authors themselves as authors often reply to other commenters and start threads. This just leaves kudos. Kudos are at the bottom of each chapter of a work, so while it at least partially guarantees the person read at least one chapter it doesn’t guarantee multi-chapter works. However, people usually leave kudos when they really like a work as it is tied to their account unless they aren’t logged in. After that I scraped the top 2,500 works of this fandom using python. I then used excel to clean up the data and assign work IDs in order to load the data into Gephi. Gephi was used to analyze how relationships interact with each other and what minor characters were tied to major characters like Alex, Kara, and Lena.
After that I used docuscope to generate text CSV files from the corpus of fanfictions that were downloaded. This allowed Docuscope to read through the files and analyze them. Additionally, Docuscope was used to analyze one of the highest fanfictions on the negativity category. The negativity category references negative acts, emotions, relationships, and values. The CSV files that were generated using Docuscope were then uploaded to Tableau, where I generated several graphs for analysis. I generated a bar graph to get an overview of the minor characters. Additionally, I looked at the number of works that had alternate universes as opposed to canon works. From those graphs I then looked at the negative language and positive language used on average for both of these tags. I also generated a graph to see on average whether minor characters were associated more commonly with canon poaching or AU poaching. Lastly, I generated a graph of characters from other fanfictions that were played by the same character to see how common it was.
In addition to data analysis, I would also like to gain some perspective from fanspaces. The method I chose to do this through is Reddit since you can search for specific posts on the sight. Supergirl’s subreddit is still relatively active, but I would like to go through old posts to see some fan favorite minor characters and see what their thoughts are on certain plot points that are contentious in fanfiction works found throughout the data analysis.
