I was trying to analyze a lot of data from my website — GA4, user events, user activity, payments, and other product data. When I tried uploading large tables directly to Claude, it didn't work well. I also didn't want to keep writing and running the same data science algorithms every time I wanted to answer a new question. I looked at some existing open-source alternatives, but most of them mainly perform SQL queries and basic analysis on the data. I wanted to go beyond that. I wanted to ask questions like: * What factors are correlated with user churn? * Which users are likely to churn? * Can we predict future behavior? * What are the important features affecting a particular outcome? So I built this MCP server to combine regular data analysis with actual data science algorithms such as feature extraction, classification, clustering, and prediction. The goal was to make it easier to go from raw data to useful insights without having to write the same data science code over and over again.

GitHub

MCP server for data analysis

MCP server for data analysis