🌠Deep Data

An agent framework built on Claude SDK for data analysis, visualization, and ML automation.

DeepData.mp4

Features

1. Agent - Coding agent with Visualization State API (IVG) for interactive chart creation and verification

   

2. Deep Plot - Autonomous data analysis agent with iterative exploration and explanation

   

3. MLE - MCTS-based ML solution search with parallel workers (isolated in Git worktrees)

   

Requirements

• Python: 3.10+
• Claude API: Via Claude Code CLI

Installation

pip install claude-deepdata

Quick Start

# Start web server
deepdata

# With custom port
deepdata --port 8080

# Set working directory
deepdata --cwd /path/to/project

Visit http://localhost:8000 for the Web UI.

MLE Environment Setup

The MLE feature requires a separate ML environment with additional dependencies:

pip install -r requirements_ml.txt

Development Setup

For contributors who want to modify the code:

# Clone the repository
git clone https://github.com/YiyangLu/Deep-Data.git
cd Deep-Data

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install in editable mode with dev dependencies
pip install -e ".[dev]"

# Build frontend (requires Node.js 18+)
cd deepdata/web/frontend
npm install
npm run build
cd ../../..

# Run the server
deepdata

Paper

Our paper describing the IVG framework benchmark is coming soon on arXiv.