LLMs in R for Data Analysis

rainbowR Conference Workshop

by Nic Crane

Wednesday 25th February 2026, 19:00-21:00 UTC / 14:00-16:00 ET / 11:00-13:00 PT

Workshop Overview

LLMs can be massively useful in R workflows, when used in the right places and if you’re aware of not only the benefits, but also the risks. In this hands-on workshop, you’ll learn how to use LLMs programmatically in R, and come away with both the confidence to experiment and a working script that extracts data from unstructured text.

We’ll look at where LLMs can help, where they’ll let you down, and techniques for making your results more trustworthy. You’re going to leave with a practical sense of what’s possible and where to be cautious.

What you’ll learn:

Getting started with LLMs in R using {ellmer}
Prompt engineering for more predictable results
Extracting structured data from unstructured text
Using LLMs to call R functions (tool calling)

Workshop Prework

Environment

You’ll need an R environment to work in; this could be RStudio or Positron on your own laptop, or an account on Posit Cloud via a web browser.

If you’re on a work machine with restrictions on installing software or connecting to external APIs, Posit Cloud is a good option as everything runs in your browser.

Packages

Install the required packages:

install.packages(c("ellmer", "dplyr", "readr", "tidyr", "usethis"))

API keys

You’ll need an API key from an LLM provider. You will have been sent this on the day of the workshop.

Workshop Schedule

Section 0: Welcome and Setup
Section 1: Getting Started with LLMs in R
Section 2: Prompt Engineering
Section 3: Structured Output
Section 4: Tool Calling
Section 5: Wrap-up and Resources

Instructor

Nic Crane is an R consultant with a background in data science and software engineering. They are passionate about open source, and learning and teaching all things R.

Code

All the code from the workshop can be found on the Code and Exercises page.