A GitHub-ready project that fetches business reviews from Google Maps (scrape first, API fallback), surfaces top keywords, and runs aspect-based sentiment analysis with a React dashboard.

• Business name or Google Maps URL input
• Scrape Google Maps reviews (Playwright)
• Fallback to Google Places API when scraping fails (if API key present)
• Auto keyword/topic suggestions when user doesn’t provide aspects
• Aspect-level sentiment analysis
• Interactive charts and review explorer

• backend/ FastAPI service (scraping + analysis)
• frontend/ React (Vite) dashboard
• *.ipynb Original notebooks kept for reference

• Python 3.10+
• Node 18+

cd backend
python -m venv .venv
.\.venv\Scripts\activate
pip install -r requirements.txt
python -m playwright install

Create backend/.env based on backend/.env.example if you want API fallback. You can also set SCRAPE_FIRST=false to skip scraping and go straight to the API.

The scraper is now limited to Google Travel hotel review URLs, but you can also paste a hotel name and it will auto-navigate to the first Google Travel result. Paste a hotel review URL like: https://www.google.com/travel/hotels/entity/.../reviews?...

Run:

uvicorn main:app --reload --port 8000

cd frontend
npm install
npm run dev

The frontend proxies /api to http://localhost:8000 in dev.

Google Maps DOM changes frequently, so scraping can break. The backend always tries scraping first, then uses Places API if a key is available.

• Add caching and rate limiting
• Persist analysis results
• Add user accounts