🎭 AI Sentiment Analyzer

A full-stack web application for real-time and batch sentiment analysis. Built with React + Vite frontend and a Flask backend, integrated with Hugging Face Transformers, Scikit-learn, and a heuristic rule-based model.

📌 Features

🔹 Frontend (React + Vite)

Responsive, modern UI (textarea, charts, file upload)

Real-time single review analysis with delay/debounce

Batch CSV upload and analysis (counts, keywords, charts)

Model selector (Hugging Face / Scikit-learn / Heuristic)

🔹 Backend (Flask)

RESTful API with /api/health, /api/predict, /api/batch, /api/debug

Hugging Face cardiffnlp/twitter-roberta-base-sentiment-latest

Scikit-learn: TF-IDF + Logistic Regression (toy dataset)

Heuristic rules + keyword triggers for fallback/edge cases

Batch CSV parsing with keyword summary & counts

🔹 Models

Hugging Face: Pre-trained transformer (POS/NEU/NEG)

Scikit-learn: Lightweight ML baseline

Heuristic: Simple keyword-based rules with overrides

⚙️ Installation & Run 1️⃣ Backend (Flask) cd backend python3 -m venv .venv1 source .venv1/bin/activate

Install dependencies

pip install -r requirements.txt

Run server

export API_TOKEN=dev-token python3 app.py

Runs at → http://127.0.0.1:5000

2️⃣ Frontend (React + Vite) cd frontend npm install npm run dev

Runs at → http://127.0.0.1:5173

🔗 API Endpoints Health curl http://127.0.0.1:5000/api/health

→ {"status":"ok"}

Predict curl -X POST http://127.0.0.1:5000/api/predict
-H "Authorization: Bearer dev-token"
-H "Content-Type: application/json"
-H "X-Model: hf"
-d '{"text":"I love this!"}'

Batch printf 'text\nLove it\nWorst phone\nOkay product\n' > reviews.csv

curl -X POST "http://127.0.0.1:5000/api/batch?col=text"
-H "Authorization: Bearer dev-token"
-H "X-Model: sk"
-F "file=@reviews.csv"

📊 Screenshots

🧠 Models

Hugging Face cardiffnlp/twitter-roberta-base-sentiment-latest → Pre-trained 3-class transformer

Scikit-learn TF-IDF + Logistic Regression → Trained on toy dataset (30 samples)

Heuristic Rule-based triggers (love, hate, okay, etc.) → Fast fallback model

🚀 Future Improvements

Hybrid model: Hugging Face embeddings + Scikit-learn classifier

Train and save/load real Scikit-learn models (joblib)

Loading spinners & error toasts in frontend

Docker + AWS deployment for public demo

Multi-language sentiment support

📂 Project Structure ai-sentiment-analyzer/ ├── backend/ # Flask API + models │ ├── app.py │ ├── requirements.txt │ └── ... ├── frontend/ # React + Vite UI │ ├── src/ │ ├── package.json │ └── ... └── README.md