Ace Attorney Project

This project is based on the dialogue from the Ace Attorney visual novel franchise, developed and published by Capcom.
The series first released in 2001, on the Gameboy, but since then, the series has had 8 major installments and 3 spinoff titles, and between rereleases and bundles, the franchise is playable on every modern major console.
As you can imagine, the franchise has a ton of text, text is what half of the entire game genre is composed of. What data could you glean from all that text?
Additionally, what would a text-generative model look like trained entirely on the games' text? Could AI hold a candle to one of the most prestigious visual novels out there? Court is in session!
Here's a short guide to the pages on this site:

• Data: Take a look at the data we gathered in the project's lifetime
• Analysis: Check out visualizations of the structured data created from the games' text
• AI-Generated Cases: Explore some text generated by models trained using the franchise's content
• Members: Introduces each member of the team

This work is marked with CC0 1.0