GitHub Personality Analyzer

Enter any GitHub username and get a "developer personality" report: profile stats, personality badges, language breakdown, a commit heatmap, top repos, and an AI-written summary (powered by Google's Gemini free tier). Includes a compare mode for two developers side by side.

✨ Recently added

• Light/dark theme toggle — switch themes from the header button; preference is remembered (localStorage) and respects the system theme on first visit. The whole palette is CSS-variable driven so all existing cards/charts adapt.
• Export report as image — "🖼️ Export as image" renders the profile card, badges, charts, and repo list to a downloadable PNG via html2canvas.
• Copy as Markdown — alongside the plain-text share summary, a "📄 Copy as Markdown" button copies a README-ready snippet with badges and stats.
• Per-repo personality tags — each top repo now shows tags like "📖 Well documented", "🚀 Popular", "⚡ Actively maintained", "🕸️ Stale", or "🍴 Fork magnet", computed client-side from data already returned by the API.
• Streak badges — the commit heatmap's longest-streak calculation now feeds a badge ("🔥 Consistent Committer" / "On a Streak" / "Unstoppable" at 7/14/30+ days) instead of only showing up in the heatmap tooltip.
• Animated three.js background — an ambient constellation/graph of drifting, connecting nodes rendered behind the whole app in the GitHub blue/green palette. Subtle mouse parallax, pauses on hidden tabs, and respects prefers-reduced-motion.
• Commit Rhythm chart — hour-of-day and day-of-week bar charts built from the push-event time distribution the backend already computed but the UI never showed.
• Personality Radar — a 6-axis radar chart (open source score, docs, night owl, early bird, weekend activity, bug hunting) for an at-a-glance shape of a profile.
• Recent searches — the last 8 usernames you looked up are saved locally (localStorage) as quick-access chips under the search bar.
• Copy shareable summary — one click copies a plain-text recap of the report (stats, top languages, badges, scores) to the clipboard.
• Subtle fade-in-up entrance animation on report cards.

Stack

• Frontend: React (Vite) + Tailwind CSS + Chart.js
• Backend: Node.js + Express
• APIs: GitHub REST API + GitHub Search API, Gemini API (gemini-2.0-flash)
• Caching: in-memory (node-cache), 15 min TTL, to stay within GitHub's rate limits

Setup

1. Backend

cd backend
cp .env.example .env

Edit .env:

PORT=5000
GITHUB_TOKEN=your_github_personal_access_token   # optional but recommended, no scopes needed
GEMINI_API_KEY=your_gemini_api_key                # required for AI summary/roast
GEMINI_MODEL=gemini-2.0-flash
CLIENT_URL=http://localhost:5173

• Get a GitHub token (classic, no scopes) at https://github.com/settings/tokens — this raises your GitHub API limit from 60/hr to 5000/hr. Without it the app still works for light usage.
• Get a free Gemini API key at https://aistudio.google.com/app/apikey.

npm install
npm run dev      # nodemon, or `npm start` for plain node

Backend runs on http://localhost:5000.

2. Frontend

cd frontend
npm install
npm run dev

Frontend runs on http://localhost:5173 and proxies /api calls to the backend.

API endpoints

• GET /api/github/analyze/:username — full profile + stats + badges + languages + heatmap
• GET /api/github/analyze/:username?deep=true — same, but pulls real per-repo byte counts for languages (slower, more accurate)
• GET /api/github/compare/:userA/:userB — both users' analysis in one response
• GET /api/ai/summary/:username — Gemini-generated personality summary
• GET /api/ai/roast/:username — optional light-hearted roast
• GET /api/ai/compare-verdict/:userA/:userB — Gemini verdict comparing two developers

Notes on data accuracy

• Commit heatmap / night-owl / early-bird badges are derived from the GitHub Events API, which only exposes a rolling ~90 days of public activity (a hard GitHub limitation, not configurable). This is a good sample for personality inference, but not a full history.
• Bug fix ratio uses the Search API (author:<user> type:pr label:bug) as an approximation — it depends on repos actually using a "bug" label.
• Language breakdown defaults to counting each repo's primary language (fast, low API usage). Pass ?deep=true to fetch real byte-level breakdowns for the top 15 starred repos instead.

Deploying (matches a typical PM2 + Nginx setup)

# Backend
cd backend && npm install --production
pm2 start server.js --name github-personality-api

# Frontend
cd frontend && npm install && npm run build
# serve the `dist/` folder via Nginx, proxying /api to the PM2 process

Roadmap ideas not yet built

• Persisting analyses to MongoDB for history/leaderboards
• Redis caching layer instead of in-memory cache (for multi-instance deployment)
• GraphQL-based full contribution calendar (needs OAuth or a broader-scoped token)