📊 EDA Autogen App

EDA Autogen App is an advanced, interactive tool for automatic Exploratory Data Analysis (EDA), powered by multi-agent AI and flexible LLM support. Users can upload CSV datasets and watch AI agents collaborate to analyze data, generate Python code, review outputs, and provide actionable insights—all via a Streamlit web interface.

---

🚀 Features

• ✅ Modern Web Interface — Responsive Streamlit UI with drag & drop CSV upload.
• 🤖 Multi-Agent Collaboration:
  • User Agent: Orchestrates the EDA process.
  • Coder Agent: Generates Python code for analysis and visualization.
  • Critic Agent: Reviews code quality and suggests improvements.
  • Analyst Agent: Provides business insights and recommendations.
• 📊 Live Code Execution — Python code from agents is executed in real time, including plots and DataFrame modifications.
• ⚙️ Configurable Parameters — Select LLM provider, model, creativity, and conversation depth.
• 🔒 Flexible LLM Support — Works with Ollama (local), OpenAI, or Gemini (cloud).
• 💾 Exportable Reports — Download full analysis including all agent messages and visualizations.
• 🧠 Auto-Generated Insights — Comprehensive analysis with actionable recommendations.
• 🛠️ Performance Tracking — Progress bars, execution time estimates, and error warnings.

---

🧩 Workflow

1. Upload Dataset

   • Drag & drop CSV file in Streamlit UI.
   • Preview basic statistics: rows, columns, data types, memory usage.

2. Configuration

   • Select LLM provider: Ollama (local), OpenAI, or Gemini (cloud).
   • Choose model and adjust temperature & max conversation rounds.
   • Provider-specific status and requirements displayed.

3. Start Analysis

   • Initializes multi-agent group chat (User, Coder, Critic, Analyst).
   • Passes dataset and configuration to agents.

4. Agent Collaboration

   • Agents exchange messages to perform EDA.
   • Coder generates Python code, Critic reviews it, Analyst provides business insights.
   • All interactions displayed live in the UI.

5. Live Execution

   • Code executed on the fly, plots rendered via matplotlib/seaborn.
   • Data modifications and errors shown in real time.

6. Results & Export

   • Display of summary stats, plots, agent conversation.
   • Downloadable full analysis report.

---

🛠️ Architecture & Tech Stack

Layer	Technology/Module	Purpose
Frontend	Streamlit	Interactive web UI, file upload, visualization
Agents	eda_agents.py	Multi-agent orchestration (User, Coder, Critic, Analyst)
LLM Client	llm_client.py	Unified interface for Ollama, OpenAI, Gemini
Config	config.py	Provider/model selection, API keys, defaults
Data	pandas, numpy	Data loading and processing
Visualization	matplotlib, seaborn	Plotting and charts
Other	requests, time, os, sys	API calls, performance tracking

---

🤖 LLM Provider Support

• Ollama: Local and offline, supports models like mistral, llama3, codellama, qwen. Privacy-focused.
• OpenAI: Cloud-based GPT models, requires API key (GPT-3.5, GPT-4).
• Gemini: Google cloud LLM, requires API key, supports Gemini-1.5 models.

The unified client (llm_client.py) abstracts differences between providers for seamless switching and robust error handling.

---

🐛 Troubleshooting & Extensibility

• Provider not configured: UI provides instructions to start Ollama or set API keys.
• Add new agents/models: Extend eda_agents.py and update config.py.
• Custom workflows: Modify agent prompts and system messages.

---

📦 Project Structure

eda-autogen-app/
├── app.py                 # Main Streamlit app
├── run.py                 # Launcher script
├── requirements.txt       # Dependencies
├── backend/
│   ├── agents/            # Agent definitions
│   │   ├── eda_agents.py
│   │   ├── coder.py
│   │   ├── critic.py
│   │   └── user_proxy.py
│   └── llm_client.py      # Unified LLM client
├── config.py              # Provider/model config
├── .gitignore
├── README.md
└── LICENSE

🚀 Getting Started

Prerequisites

• Python 3.8 or higher
• Ollama (for local LLM)
• API keys for OpenAI or Gemini (if using cloud LLMs)

Installation

# Clone the repository
git clone (https://github.com/Narendersingh007/eda-autogen-app)
cd eda-autogen-app

# Install dependencies
pip install -r requirements.txt

### Launch App

```bash
# For local LLM
ollama serve
ollama pull mistral

# Set API keys (if using cloud LLMs)
export OPENAI_API_KEY='your-openai-key'
export GEMINI_API_KEY='your-gemini-key'

# Start Streamlit app
streamlit run app.py

---

🌟 Why This Project Stands Out

• Combines multi-agent AI with real-time code execution and visualization.
• Supports both local and cloud LLMs for flexibility, privacy, and performance.
• Professional UI/UX with live feedback, error handling, and progress tracking.
• Extensible architecture allows adding new agents, models, and workflows.
• Ideal portfolio project demonstrating Python, Streamlit, AI/ML, and full-stack integration skills.

---

👤 Author

Narender Singh
✉️ narendersingh123987@gmail.com
🔗 GitHub Profile

---

📜 License

This project is licensed under the MIT License — see the LICENSE file for details.

---

💬 Contributing

Contributions are welcome! Follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/your-feature)
3. Make your changes
4. Add tests if applicable
5. Submit a Pull Request

---