Cursor is a code editor built on VS Code with AI chat deeply integrated — not just autocomplete, but a full conversation about your code and your codebase. For researchers who write Python or R for data analysis, it removes the context-switching between writing code and asking ChatGPT about code.
Why Cursor Over Copilot for Research
Copilot gives inline completions as you type. Cursor gives both inline completions AND a chat panel where you can ask questions about files you have open, ask it to rewrite functions, debug errors by pasting them into chat, or ask “explain what this code does” on any selected block. The chat has context of your entire project — it can see all your files and answer questions across them.
Setup for Python Data Analysis
Install Cursor (cursor.sh), sign in with GitHub, and open your analysis project folder. Configure the AI model to Claude or GPT-4 in settings. Create a `.cursorrules` file in your project root describing your project: “This is a Python data analysis project using pandas, scikit-learn, matplotlib. Data is loaded from the /data directory. Analysis scripts are in /analysis. We follow APA style for figure formatting.” The AI now has project-specific context for all suggestions.
Practical Workflows
Debugging: paste the error message into Cursor chat with the question “What causes this error and how do I fix it?” It reads the error, finds the relevant lines in your code, and proposes a fix. Faster than Stack Overflow for common errors.
Refactoring: select a messy data cleaning function and press Cmd+K (inline edit). Type “make this cleaner and add descriptive variable names.” Cursor rewrites the selected block in place.
Documentation: select a function and ask “Write a docstring for this function explaining what it does, its parameters, and what it returns.” Essential for returning to analysis scripts after 6 months away.
R Users
Cursor supports R. The setup is the same. Configure your `.cursorrules` to mention R, tidyverse, ggplot2, and your statistical context. The AI understands tidyverse idioms and ggplot2 layer syntax, which makes it more useful than generic AI assistants for R-specific questions.
Pricing
Cursor has a free tier with limited AI usage per month. The Pro plan is $20/month with unlimited AI. For a PhD student writing data analysis code daily, Pro is worth it. There are student discounts; check cursor.sh/pricing.
