Online creative writing platforms provide a useful setting for studying how users discover content, form communities, and respond to different types of writing. This project began as an investigation of engagement patterns across multiple creative writing platforms, with a focus on the relationship between emotional tone, metadata, and reader response.
For the main analysis, I selected Archive of Our Own (AO3) because of its unusually flexible and detailed tagging system. AO3 allows users to describe works through fandom, relationship, character, and freeform tags, making it especially well suited for studying how metadata and community structure relate to content discovery and engagement.
I used hits as a measure of popularity and kudos per hit as a normalized measure of likability, allowing me to distinguish between works that receive broad exposure and works that receive stronger positive engagement from readers.
I applied hedonometric and ousiometric sentiment analysis to story text to quantify emotional tone across the dataset. I then aggregated these results at the tag and fandom levels to examine relationships between sentiment and engagement.
To study tagging behavior, I developed fandom-specific tag-popularity association scores that compare the popularity of works containing a given tag with other works in the same fandom. This made it possible to measure how the relationship between tags and engagement varies across different communities.
I also constructed a weighted fandom co-occurrence network using associated works and compared network structure with fandom-specific tag behavior. The analysis showed a positive relationship between network distance and differences in tag-popularity patterns, suggesting that more closely connected fandoms tend to exhibit more similar engagement dynamics.
Overall, the project combined natural language processing, statistical analysis, and network theory to study how emotional tone, metadata, and community structure are associated with content discovery and user engagement in online creative writing communities.