Add tutorial for analyzing data with Ghost and an AI agent

[akulkarni]

Summary

• New tutorial at docs/tutorials/movielens-analysis.md walking through end-to-end MovieLens data analysis using Ghost plus an AI coding agent (Claude Code, Cursor, Codex, Windsurf, etc.).
• Structured for agent-driven execution: a one-time human bootstrap (install Ghost, login, install MCP, restart agent), then each subsequent step is a copy-pasteable prompt the reader pastes into their agent — download the dataset, install psql if missing, create the database and load CSVs (with the agent inferring the schema), ask three analytical questions, delete when done.
• Each agent prompt is self-contained: platform branches and fallbacks (e.g. unzip vs. python3 -m zipfile, per-OS psql install) live inside the prompt block so the prompt works without surrounding context.

Test plan

• Read top-to-bottom from a fresh checkout to verify the flow makes sense.
• Run the tutorial through with Claude Code on macOS — confirm Ghost CLI install, MCP install, and each agent step works as written.
• Spot-check on Linux (Debian/Ubuntu) — confirm the psql and unzip install branches in the agent prompts behave as expected.
• Confirm the three analytical query results land roughly where the expected-output examples claim.
• Verify the resource links resolve (Ghost docs, MovieLens, MCP install reference).

🤖 Generated with Claude Code