Overview
Data Security
Approaches
Writing Code
Analyzing Data
Next Week
AI has gotten amazing, but you still need to verify results
Why learn to code? Because foundational skills let you do more
Many approaches, but Claude Code is the best right now
Access to data is central to our tools’ ability to do their jobs. It is difficult to successfully anonymize information that’s passed to an LLM, especially if that LLM has access to execute arbitrary code. - Simon Couch and Sara Altman
Using an AI data science agent while obscuring your data is similar to hiring a financial advisor but hiding the contents of your bank accounts. If you don’t trust a financial advisor to look at your accounts, they will be of little use to you and you probably shouldn’t hire them. - Simon Couch and Sara Altman
If you have super sensitive data, don’t use AI tools
Set your account so chats are not used to train models
Know that your data will leave your computer (unless using a locally-run LLM)
Review the privacy policies of LLM providers to find one you’re comfortable with
Web UI
Desktop app (e.g. Claude)
In Positron or RStudio (via Positron Assistant / Posit Assistant)
IDE extension
Terminal (Claude Code)
Bouncing between a chatbot and RStudio feels like the old days of copying and pasting plots into a Word document. And because the chatbot doesn’t know what’s in your R session, it often gives the wrong answer.
— Hadley Wickham
Posit AI closes the loop by embedding the chatbot in RStudio, letting it see what data and packages you have loaded and run code on your behalf.
— Hadley Wickham
Code completion
Fix and Explain
Agentic AI
Positron Assistant
Claude Code
Positron Assistant
Using GitHub Copilot
Using Claude
Claude Code
Posit Assistant
Plan
Ask before edit
Edit automatically
Bypass permissions
Write code to analyze data
Databot
Course assignment: advanced data wrangling, part 1 (tidy data)
No project assignment in week 6