Studio Ghibli films are trending and beginning to capture many peoples hearts — but what new stories do fans secretly wish existed?

I worked on a project that uses machine learning to explore that question, the purpose was to uncover emerging storytelling trends and help studios predict what kinds of content audiences want next.

In this project, I analyzed a synthetic dataset of 500+ AI-generated Ghibli-style images using techniques from:
- Natural Language Processing (TF-IDF, LDA Topic Modeling)
- Supervised Learning (Random Forest Classifier)
- Feature Engineering & Sentiment Analysis (VADER)
- Exploratory Data Analysis (Seaborn, Matplotlib)

I clustered text prompts to uncover high-potential themes like rural escapism and post-apocalyptic hope, then trained a model to predict high-engagement art. The goal? Help animation studios use AI-generated prompts as crowdsourced story forecasts, reducing creative risk and building content people already love.

Bonus: Built a Streamlit dashboard to visualize prompt trends and audience preferences in real time.