E_LENS — E-commerce Analysis AI Agent

AI-powered e-commerce data analysis system built with LangGraph, LangChain, and Google Gemini.

E_LENS turns natural-language questions into data analysis, visualizations, insights, and reports through a multi-agent workflow.

Features

• 🤖 Multi-agent AI analysis
• 📊 Automated EDA, correlation & feature analysis
• 🔎 SQL-based analysis workflows
• 💡 AI-generated insights
• 📈 Visualization generation
• 📝 Automated PDF reports
• 👤 Human-in-the-loop approvals
• 🔍 LangSmith tracing & session logging

Architecture

User
  ↓
Orchestrator
  ↓
┌─────────────┬─────────────┬─────────────┐
│ FE Agent    │ SQL Agent   │ Insight Agent│
└─────────────┴─────────────┴─────────────┘
                 ↓
           Report Agent
                 ↓
          Final Analysis

---

Tech Stack

• AI: LangGraph, LangChain, Google Gemini, LangSmith
• Data: Pandas, NumPy, scikit-learn, LightGBM
• App: Python, Streamlit, FastAPI
• Storage: SQLite, SQLAlchemy
• Reports: ReportLab, WeasyPrint

---

Run

git clone https://github.com/AdityaGit96/Ecom-Analysis-AI-Agent.git
cd Ecom-Analysis-AI-Agent

pip install -r requirements.txt

python main.py --file data/sample/ecommerce_sample.csv

---

Example

"Find variables most correlated with revenue"

• → Analyze data
• → Calculate correlations
• → Return results
• → Generate insights

---

Project Structure

src/
├── agents/
├── tools/
├── streamlit/
├── graph.py
├── state.py
└── human_in_the_loop.py

---

License

MIT