Open-source analytics for Japan’s Numbers3, Numbers4 & Loto6: historical draws, ensemble ML forecasts, SQLite + GitHub Actions pipelines, and a Next.js dashboard (optional Supabase). Made for transparency and reproducible research—not betting advice, not affiliated with any official lottery operator, and not a promise of results. Feedback from data/ML folks especially welcome.
No reviews yetBe the first to leave a review for takarakuji-ai
Maker
📌
Hi Product Hunt 👋 I’m [Takeshi Kubokawa / @kubocchi.studio], maker of Million
Pocket Orchestra.
I started this project because I wanted a place where lottery draw
data isn’t just “numbers on a page,” but something you can
explore with clear pipelines: historical results, ensemble ML
forecasts, charts, and reproducible workflows (SQLite + GitHub
Actions) instead of a black box.
It focuses on Japan’s public draws (Numbers3, Numbers4, Loto6) and
ships with a Next.js dashboard (optional Supabase) plus Python
tooling. Everything is open source so you can verify how
predictions are produced and extend the models yourself.
Important context: this is for research, learning, and
transparency—not gambling advice, not affiliated with any official
operator, and not a guarantee of outcomes.
🔗 Source: https://github.com/kubokawa-dev/...
I’d love your thoughts—especially if you’re into data science or
OSS: what would you want to see next (better explainability, more
benchmarks, docs for contributors)?
Thanks for checking it out, and happy to answer questions in the
thread!