AI Agent for E-commerce Data

A powerful AI-driven web application that allows users to query e-commerce data using natural language. The system converts natural language questions into SQL queries, executes them against a database, and provides beautiful visualizations of the results.

📌 Tip: Want to see it in action? Check out the live demo GIF below!

---

✨ Features

• 🧠 Natural Language Processing: Ask questions in plain English about your e-commerce data
• 🪄 AI-Powered SQL Generation: Automatically converts questions to SQL queries using LLM (with fallback patterns)
• 🌈 Beautiful Web Interface: Modern, responsive React frontend with Tailwind and shadcn/ui
• 📊 Data Visualization: Auto-generates visual insights from query results
• 💬 Real-time Streaming: Typing animation for LLM responses
• 📂 Multiple Data Sources: Ad sales, total sales, eligibility data supported
• 🔗 RESTful API: Clean endpoints for integration into other systems

---

🏗️ Architecture

• Frontend: React (Vite) + Tailwind CSS + shadcn/ui
• Backend: Flask REST API with CORS support
• Database: SQLite with preloaded data
• AI Integration: Local LLM via Ollama (fallback logic included)
• Visualization: Matplotlib/Seaborn for charts

🧠 System Architecture

---

🚀 Quick Start

Prerequisites

• Python 3.8+
• Node.js 16+
• npm or pnpm

Installation Steps

# 1. Extract the zip file
cd ai_ecommerce_agent

# 2. Backend setup
pip install flask flask-cors pandas matplotlib seaborn

# Optional: Install Ollama for LLM
# https://ollama.com/download
# ollama pull llama2

# 3. Set up the database
python database.py

# 4. Frontend setup
cd frontend
npm install  # or: pnpm install

Running the App

# 1. Backend
cd ../  # Root directory
python api.py
# API will be at http://localhost:5000

# 2. Frontend (in new terminal)
cd frontend
npm run dev
# Frontend will be at http://localhost:5173

---

🌐 Usage

Web Interface

• Query Interface: Ask questions, visualize answers, get live responses
• Example Queries: One-click examples and schema info

Example Natural Language Queries

• "What is the total sales?"
• "Calculate the RoAS (Return on Ad Spend)"
• "Which product had the highest CPC?"
• "Show me products not eligible for advertising"
• "Top 10 products by impressions"

---

📡 API Endpoints

Endpoint	Method	Description
/	GET	API home
/health	GET	Health check
/query	POST	Submit a natural language query
/stream_query	POST	Real-time typing stream query
/schema	GET	Get DB schema

Example cURL

curl -X POST -H "Content-Type: application/json" \
  -d '{"question": "What is the total sales?", "visualize": true}' \
  http://localhost:5000/query

---

🧾 Data Schema

Ad Sales Table

• date, item_id, ad_sales, impressions, ad_spend, clicks, units_sold

Total Sales Table

• date, item_id, total_sales, total_units_ordered

Eligibility Table

• eligibility_datetime_utc, item_id, eligibility, message

---

🛠️ Troubleshooting

Issue	Fix
Module not found	pip install -r requirements.txt
DB not found	Run python database.py
API not connecting	Ensure Flask is running at :5000 with CORS
Frontend crashes	rm -rf node_modules && npm install
Ollama memory issue	Use fallback mode (see terminal logs)

---

📁 Project Structure

ai_ecommerce_agent/
├── data/                          # CSV data files
├── frontend/                      # React UI
│   ├── src/components/            # UI Components
│   ├── App.jsx                    # Main App
│   └── App.css                    # Styling
├── database.py                   # DB init
├── llm_interface.py              # AI logic
├── api.py                        # Flask API
├── requirements.txt              # Python deps
├── README.md                     # You're here ✨

---

🔧 Dev Guide

• Add new fallback prompts → llm_interface.py
• Add API routes → api.py
• Add UI components → frontend/src/components
• Add new data → database.py and data/

---

📄 License

This project is open-source and available for educational/demo purposes.

---

🤝 Support & Contributions

• Star the repo ⭐ if it helped you!
• PRs welcome for new features or bugfixes 🙌
• For issues, open GitHub Issues or reach out!

---