Huge Open-Source Anime Dataset: 1.77M Users & 148M Ratings

Hey everyone, I’ve published a freshly-built anime ratings dataset that I’ve been working on. It covers 1.77M users, 20K+ anime titles, and over 148M user ratings, all from engaged users (minimum 5 ratings each).

This dataset is great for:

• Building recommendation systems
• Studying user behavior & engagement
• Exploring genre-based analysis
• Training hybrid deep learning models with metadata

🔗 Links:

• Kaggle Dataset: https://www.kaggle.com/datasets/tavuksuzdurum/user-animelist-dataset (inference notebook available)

• Hugging Face Space: https://huggingface.co/spaces/mramazan/AnimeRecBERT

• GitHub Project (AnimeRecBERT Hybrid): https://github.com/MRamazan/AnimeRecBERT-Hybrid