5 and 11 November 2025

Large Language Models (LLMs) are becoming increasingly adept at writing, explaining, and debugging R code; and new tools that incorporate these models directly into our development environments are constantly being developed and upgraded. 
This hands-on workshop introduces the modern ecosystem of LLM-based tools available for R. Participants will learn to incorporate these tools into their data processing, visualization, or analysis pipelines in R for smooth and productive workflows.  

Participants should have basic R programming knowledge. No prior knowledge of LLMs is required.

By the end of the workshop, participants will be able to:

• Integrate LLM-based tools into R-based data analysis pipelines
• Generate, edit, document, and troubleshoot R code quickly and accurately
• Clean, transform, and visualize data with assistance from LLMs
• Interact with model APIs from the R console and their preferred IDE

Day 1 – Introduction and setup- November 5th;  1-6 PM Berlin time

1.    Welcome & Orientation

o    Use-case tour of what LLM-based tools can help us do 
o    Introduction to Language Models and Large Language Models
o    Responsible use and AI guidelines
o    Overview of cloud-based and local models
o    Overview of model providers
o    Pricing considerations

2.    Setup

o    Introduction to APIs 
o    Setting up API keys and managing secrets
o    Configuration and troubleshooting

3.    Tools and integrations

o    Avoiding context switching for more productivity
o    Interacting with LLMs beyond the web browser
o    IDE tools and assistants
o    RStudio addins
o    Context-aware assistants that understand R scripts and objects
o    using LLM extensions in Positron (Positron Assistant, Continue, Codeium, Cline)

Day 2  – Productivity with LLM-based tools - November 11th; 1-6 PM Berlin time

4.    LLM-assisted coding

o    Writing and documenting code with model assistance
o    Create and edit data visualization code using natural language prompts
o    Using LLMs to create and style Quarto documents
o    Context-aware coding with `gander`

5.    Extracting data from PDFs with structured output

o    Using `ellmer` to extract information from PDF files into usable data frames, including OCR (optical character recognition) for raw scans. 

6.    Model Context Protocol (MCP) tools for working with coding agents

o    Using coding agents (Positron Assistant, Claude Desktop, Cline) that can create and edit files, run code, and explore the R environment as needed.

7.    Wrap-up