Overview

Data Science Ai MCP provides AI-powered tools via the Model Context Protocol (MCP).

Tools

Tool	Description
feature_importance	Rank features by estimated importance for a prediction task.
model_comparison	Compare ML models across metrics. Returns composite scores, speed
dataset_profiler	Profile a dataset: completeness, quality issues, type distribution,
correlation_finder	Compute pairwise Pearson correlations between variables. Flags strong
visualization_recommender	Recommend visualizations based on data characteristics and analysis goal.

Installation

pip install meok-data-science-ai-mcp

Usage with Claude Desktop

Add to your Claude Desktop MCP config (claude_desktop_config.json):

{
  "mcpServers": {
    "data-science-ai-mcp": {
      "command": "python",
      "args": ["-m", "meok_data_science_ai_mcp.server"]
    }
  }
}

Usage with FastMCP

from mcp.server.fastmcp import FastMCP

# This server exposes 5 tool(s) via MCP
# See server.py for full implementation

License

MIT © MEOK AI Labs