PollPulse

A real-time live polling and audience engagement platform — create a poll, share a 5-character code, and watch votes update on every connected screen instantly, no refresh needed.

Think Slido/Mentimeter, minimal version: built to demonstrate WebSocket-based real-time architecture end-to-end, not just a CRUD app with a REST-only "refresh to see results" model.

Why this exists

Most portfolio full-stack projects are static request/response CRUD apps. PollPulse demonstrates:

• Real-time push architecture — votes broadcast to every connected client over WebSocket (STOMP over SockJS), not polling the server every N seconds
• Idempotency / abuse prevention — a client-side fingerprint + a unique DB constraint on (poll_id, voter_fingerprint) stops duplicate votes without requiring login
• Clean separation of concerns — Spring Boot REST + WebSocket backend, decoupled React SPA frontend, deployable independently

Architecture

Browser A ---REST(create poll)---> [Spring Boot API] ---> [Postgres]
Browser B ---REST(vote)-----------> [Spring Boot API] ---> [Postgres]
                                          |
                                   broadcast update
                                          v
                              [STOMP /topic/polls/{id}]
                                    /          \
                                   v            v
                          Browser A (live)   Browser B (live)

• Backend: Spring Boot REST API for poll CRUD + voting, Spring WebSocket (STOMP) for live broadcast on every vote.
• Frontend: React (Vite) SPA — create/join/vote/watch-live screens, subscribes to the poll's WebSocket topic on load.
• Database: Postgres in production, defaults to in-memory H2 for zero-config local dev.

Tech stack

Java 17 · Spring Boot 3 · Spring WebSocket (STOMP/SockJS) · Spring Data JPA · PostgreSQL/H2 · React 18 · Vite · React Router

Running locally

Option A — one command (Docker)

git clone <your-repo-url>
cd pollpulse
docker-compose up --build

Frontend: http://localhost:5173 · Backend: http://localhost:8080

Option B — manual (faster iteration while developing)

# backend (uses in-memory H2, no DB setup needed)
cd backend
./mvnw spring-boot:run

# frontend (separate terminal)
cd frontend
npm install
npm run dev

API

Method	Endpoint	Description
POST	/api/polls	Create a poll: { "question": "...", "options": ["A","B"] }
GET	/api/polls/{id}	Get poll by ID
GET	/api/polls/join/{code}	Get poll by its 5-character join code
POST	/api/polls/{id}/vote	Cast a vote: { "optionId": "...", "voterFingerprint": "..." }
POST	/api/polls/{id}/close	Close a poll to new votes
WS	/ws (STOMP)	Subscribe to /topic/polls/{id} for live updates

Deployment guide (free tier friendly)

This one is genuinely easy to get live for a resume link:

1. Backend → Render: New Web Service, point at /backend, Docker runtime. Add a free Render Postgres instance and set DB_URL, DB_USER, DB_PASSWORD env vars on the service. Render gives you a public HTTPS URL automatically.
2. Frontend → Vercel or Netlify: point at /frontend, build command npm run build, output dir dist. Set the env var VITE_API_BASE_URL to your Render backend URL.
3. Update the WebSocket connection — since VITE_API_BASE_URL is reused for both REST and the /ws SockJS endpoint, no extra config needed once that env var is set correctly.

Total cost: $0 on free tiers (Render free web service + free Postgres, Vercel free hobby tier). This is the one to actually put a live demo link for on your resume/LinkedIn — it's visual, interactive, and recruiters can try it themselves in 10 seconds via the join code.

License

MIT — free to use as a reference or starting point.