StatPlay is an open educational resource for self-learners of statistics. Each topic turns a statistical concept into something you can touch — drag sliders, watch distributions reshape, and see the intuition emerge before the formula.
Site: https://statplay.sasailab.com/Source code: https://github.com/sasai-lab/statplay-opensourceLicense: CC BY-NC 4.0
The site is bilingual (Japanese / English) and works offline as a Progressive Web App.
Coverage spans 16 interactive topics and 5 long-form columns:
- Descriptive statistics: standard normal, normal distribution & standardization, and the three test distributions (t, χ², F).- Probability & inference: probability rules with Venn diagrams, Bayes' theorem, the law of large numbers (LLN), the central limit theorem (CLT), confidence intervals, hypothesis testing, proportion testing, chi-squared tests, and ANOVA.- Relational statistics: correlation, simple linear regression (OLS), and multiple regression with confounder control.- Columns: deviation values, the birthday paradox, standardization, income prediction, and Type I vs Type II errors — each pivots on a moment where intuition typically slips.
Built entirely with vanilla JavaScript and HTML Canvas (zero framework dependencies) for long-term reproducibility. Every visualization is seed-reproducible: any state can be shared via URL parameters (e.g. ?seed=12345&n=300), making StatPlay citable in papers and teaching materials.
The full source code is openly available on GitHub. Educators are free to screen-record, fork the repository, or adapt visualizations for course materials with attribution ("© 2026 Sasai Lab"). The zero-dependency architecture is designed to remain functional for over a decade without framework migration.
Concrete examples use Japanese open data — national wage census, birthday distributions — a perspective rarely available in English-language statistics education.