DuckDB is a more convenient and performant tool for analyzing large CSV/JSON files. Shameless plug: if you’re interested in combining DuckDB with Claude using the recently announced Model Context Protocol (MCP), check out the MyDuck Server project (https://github.com/apecloud/myduckserver) that I’m working on. You can follow this guide https://github.com/apecloud/myduckserver/blob/main/docs/tuto... to get started. The guide includes a conversation history showcasing how the free-tier Claude Haiku model successfully analyzed the Our World in Data Energy dataset. MyDuck Server adds a Postgres frontend for DuckDB, allowing seamless interaction with DuckDB via the official Postgres MCP server.