Stats Compass MCP server

<!-- mcp-name: io.github.oogunbiyi21/stats-compass -->

# stats-compass-mcp

Turn your LLM into a data analyst. Multiple data science tools via MCP.

Quick Start

pip install stats-compass-mcp

Claude Desktop

stats-compass-mcp install --client claude

VS Code (GitHub Copilot)

stats-compass-mcp install --client vscode

Claude Code (CLI)

claude mcp add stats-compass -- uvx stats-compass-mcp run

Note: The first connection may fail while uvx downloads the package. If this happens, disable and re-enable Stats Compass in your MCP settings — subsequent connections will be instant.

Restart your client and start asking questions about your data.

What Can It Do?

Run stats-compass-mcp list-tools to see all available tools.

How to Prompt

Start your message with "Use stats compass to..." — this tells the AI to use the Stats Compass tools instead of trying to write code or use other methods.

Use stats compass to load ~/Downloads/sales.csv and run EDA on it
Use stats compass to find my CSV files in Downloads
Use stats compass to clean the dataset and handle missing values
Use stats compass to create a histogram of the price column
Use stats compass to test if there's a significant difference in scores between group A and B
Use stats compass to train a classification model to predict churn

Tip: Without this prefix, some AI clients may try to write Python code or use shell commands instead of the Stats Compass tools — especially for tasks like finding files on your machine.

Loading Files

Local mode: Start with "Use stats compass to load..." and provide the file path or folder.

Use stats compass to load the CSV at ~/Downloads/sales.csv
Use stats compass to find my data files in ~/Documents

Remote/HTTP mode: Use the upload feature (see below).

Remote Server Mode

For Docker deployments or multi-client setups:

stats-compass-mcp serve --port 8000

File Uploads

When running remotely, users can upload files via browser:

You: I want to upload a file
AI: Open this link to upload: http://localhost:8000/upload?session_id=abc123

[Upload in browser]

You: I uploaded sales.csv
AI: ✅ Loaded sales.csv (1,000 rows × 8 columns)

Downloading Results

Export DataFrames, plots, and trained models:

You: Save the cleaned data as a CSV
AI: ✅ Saved. Download: http://localhost:8000/exports/.../cleaned_data.csv

Connect Clients to Remote Server

VS Code (native HTTP support):

{
  "servers": {
    "stats-compass": { "url": "http://localhost:8000/mcp" }
  }
}

Shortened. The full README is on GitHub.