Grade Cursor on a real data science project
A data scientist tested Cursor on a real project: weak on pandas, data cleansing and scraping, strong at prototyping.
Done withCursor
- What they did
- Mike Woodward used Cursor, with Claude as the underlying model, on a real Python data science project about English football data. He graded it A to D across general coding and specific tasks. He prompted for code, then reviewed and fixed it by hand. That included a pandas query for clubs appearing only as home or only as away, which he wrote himself.
- How it went
- Grades were A for completion, commenting, documentation and prototyping, and C for pandas, data cleansing and code generation. Scraping and getting started scored D. Streamlit apps came together fast, but anything technical needed careful checking.
- Worth knowing
- The free tier runs out quickly, so a paid subscription is needed to evaluate it. Cursor also deleted some data files during GitHub use, so keep backups.
Try it yourself with Cursor
I have a data science project in [folder path] about [goal]. Try these tasks and grade yourself on each: [pandas transformations / data cleaning / web scraping / prototyping a model]. For each, tell me what worked, what you were unsure about, and what I should verify by hand. Finish with a short scorecard, and do not overwrite my raw data files.
Source: blog.engora.com · Jun 1, 2025
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
Keep coding agents running autonomously for long stretches
Cursor's write-up on scaling long-running autonomous coding.
Run a scheduled personal assistant that builds a task dashboard
A Rippling engineer's Cursor cron agent runs every two hours, reading Slack notes, GitHub PRs, Jira issues and mentions, deduplicating them and posting a dashboard.
Migrate 180 microservices to a new framework version
Brex built Cursor-based self-correcting agents that let one engineer migrate 180 microservices to a newer Micronaut version in about a month.
Find connections across 100 non-fiction books
Claude Code browses a personal library with CLI tools and surfaces trails of related excerpts across books.
Build searchable knowledge bases from videos and papers
An OpenClaw agent runs a YouTube summariser and a paper-condensing tool, saving Markdown into two indexed knowledge bases for videos and papers.
Turn voice notes into structured Obsidian notes
Voice notes sent over Telegram become Obsidian thought notes with transcript, cleaned argument, linked context and a draft; Hermes also tracks recipes.