StatSprig is an interactive statistical test selection and reference app built to help users find a reasonable starting point for an analysis, understand the assumptions behind common methods, and compare nearby alternatives without leaving the browser.
Live site: statsprig.com
Download offline version: docs/index.html
Current library:
StatSprig is designed for students, researchers, educators, analysts, and anyone who wants a faster way to move from a research question to a defensible method choice.
The current product includes:
Most of StatSprig's decision logic and content live in typed frontend data files rather than a live backend service:
client/src/lib/wizardKeys.tscontains the decision flow and recommendation rulesclient/src/lib/statsData.tscontains the statistical test catalog and code snippetsclient/src/lib/glossaryData.tscontains glossary definitions and related terms
This keeps the app fast, portable, and suitable for both web deployment and offline usage.
Prerequisites:
- Node.js 20+
- npm
Install dependencies:
npm installStart the development server:
npm run devThis starts the app on port 5000 by default.
Useful commands:
npm run check # TypeScript type-check
npm run build # Production build + static offline exportnpm run build produces two deployment targets:
dist/for the bundled server and standard production client assetsdocs/index.htmlas a fully inlined single-file build for offline use and GitHub Pages-style static hosting
The offline build uses hash-based routing so it can run directly from file:// without a web server.
- React 18 + TypeScript
- Vite 5
- Wouter for routing, with a custom hash-router path for offline mode
- Tailwind CSS for styling and design tokens
- shadcn/ui components built on Radix UI primitives
- Lucide React icons
- Framer Motion for selected UI animation
- Typed TypeScript modules for decision rules, glossary data, and test metadata
- React Markdown for rich text rendering
remark-math+rehype-katexfor math support- Prism-based code blocks for Python and R snippets
- CSS custom properties and a custom light/dark theme toggle persisted in
localStorage
- Node.js + Express as the development and production host shell
- esbuild for bundling the server output
vite-plugin-singlefilefor the standalone offline export- A dedicated static Vite config for generating the single-file
docs/build
- The core StatSprig experience is currently client-driven
- The included server is mainly responsible for local development and production hosting
- The statistical catalog, glossary, and recommendation logic do not depend on a live database or API
StatSprig aims to make statistical method selection more practical and less opaque by emphasizing:
- Reasonable starting-point recommendations instead of false certainty
- Assumption awareness instead of memorized test names
- Quick comparison between adjacent methods
- Free access without accounts or paywalls
- A usable experience for both beginners and experienced users
- DOED Minority Science and Engineering Improvement Program (MSEIP) Grant#: P120A220015, "Boosting STEM Student Success"
- Dr. Bernadette Hence at U.S. Department of Education
- Dr. Mary Jo Parker, PI, MSEIP at University of Houston-Downtown
- Dr. Katherine Shoemaker, Co-PI, MSEIP at University of Houston-Downtown
StatSprig was developed independently, but projects such as Stat-Tree remain valuable adjacent resources for statistical test selection and statistical software walkthroughs. The broader goal is the same: make better statistical decision-making more accessible.
This repository is currently marked as MIT in package.json.