Overview
What is MCP-Server for Data Analysis?
MCP-Server for Data Analysis is an MCP server that reads Excel files from the desktop, performs descriptive analysis, and generates a variety of plots. It uses FastMCP to define and expose each tool as an endpoint, making it suitable for users who want to analyze local spreadsheet data through AI assistants like Claude.
How to use MCP-Server for Data Analysis?
Install dependencies (mcp[cli], pandas, matplotlib, openpyxl), place the server script in a project directory, and run mcp dev excel_analytics_server.py for testing. For production use with Claude Desktop, configure claude_desktop_config.json to run the server via uv.
Key features of MCP-Server for Data Analysis
- Reads .xlsx and .xls files from the desktop
- Performs descriptive statistics on numeric, categorical, and date columns
- Computes correlation matrix for numeric columns
- Creates bar, pie, line, scatter, histogram, and box plots
- Returns plots as in-memory PNG images
- Integrates with Claude Desktop via the MCP interface
Use cases of MCP-Server for Data Analysis
- Analyze sales data from an Excel file on the desktop
- Generate a bar chart of category totals from a spreadsheet
- Compute descriptive statistics for survey responses
- Create a scatter plot to explore relationships between variables
- Produce a histogram of numerical data distribution
FAQ from MCP-Server for Data Analysis
What dependencies are required?
Dependencies include pandas, matplotlib, openpyxl, and the MCP package with CLI extras.
Where are the Excel files read from?
The server scans and reads Excel files from the user's desktop directory.
How do I start the MCP server?
Run the script directly with python excel_analytics_server.py or use mcp dev excel_analytics_server.py for testing.
Can I use this server with Claude Desktop?
Yes, the README provides configuration steps to connect the server to Claude Desktop via claude_desktop_config.json.
What types of plots can be generated?
Bar chart, pie chart, line chart, scatter plot, histogram, and box plot.
