Turn reviews into insights. AI extracts what matters.

ReviewIQ is an NLP-powered web app for the AI x Data 2.0 Hackathon (Pillar 4: Make Reviews Useful). It analyzes women's clothing reviews to extract sentiment, loved features, complaints, and trending words — delivering an actionable insights dashboard.

1. Upload — Paste reviews or upload CSV
2. Analyze — Hugging Face AI processes sentiment & patterns
3. Dashboard — View charts, ranked lists, word cloud, export PDF/JSON

• Node.js 18+
• Hugging Face API token (free)

cp .env.example .env
# Edit .env with your HUGGINGFACE_API_KEY

cd backend
npm install
npm run dev
# Runs on http://localhost:3001

cd frontend
npm install
npm run dev
# Runs on http://localhost:3000

Open http://localhost:3000 → Analyze Reviews → Load sample → View dashboard.

Live demo: review-iq-smoky.vercel.app · API: reviewiq-nmak.onrender.com

Request:

{ "reviews": ["Love the soft fabric!", "Shrinks after wash..."] }

Response:

{
  "sentiment": { "positive": 72, "negative": 15, "neutral": 13 },
  "topFeatures": [{ "feature": "Soft fabric", "count": 245, "percentage": 10.6 }],
  "topComplaints": [{ "complaint": "Shrinks after wash", "count": 87, "percentage": 3.8 }],
  "wordCloud": [{ "word": "fabric", "count": 120, "sentiment": "positive" }],
  "summary": "Overall sentiment is 72% positive. Most loved: Soft fabric. Main complaint: Shrinks after wash.",
  "reviewCount": 15
}

pip install pandas
python scripts/process_reviews.py --input path/to/reviews.csv --output data/processed.json --sample 500

See DEPLOY.md for full steps.

1. Push to GitHub
2. Import repo in Vercel, set root to frontend/
3. Env: NEXT_PUBLIC_API_URL=https://your-app.onrender.com

1. New Web Service from GitHub, root backend/
2. Env: HUGGINGFACE_API_KEY=hf_... (Render sets PORT for you)
3. Build: npm install && npm run build · Start: npm start
4. Or use the included render.yaml Blueprint at repo root

• ✅ DEVPOST.md — copy-paste Devpost project page text
• ✅ README.md — setup & run instructions
• ✅ requirements.txt / package.json — dependencies
• ✅ scripts/process_reviews.py — data processing
• ✅ backend/ + frontend/ — application code
• ✅ MODEL_REPORT.md — approach & validation
• ✅ RESPONSIBLE_AI.md — ethics statement
• ✅ predictions.csv — sample analysis output

reviewiq/
├── frontend/          # Next.js app
├── backend/           # Express API + Hugging Face NLP
├── scripts/           # Data processing
├── MODEL_REPORT.md
├── RESPONSIBLE_AI.md
└── predictions.csv

MIT — built for AI x Data 2.0 Hackathon