September 2026 Meetup

An introduction to toolero

Details

• Date: Tuesday, September 1, 2026
• Time: 1:30 – 2:30 PM
• Location: CS 3139, Computer Science Building, UW–Madison

What we’ll cover

This month we’re turning to a tool a few of you have already heard me mention in passing: toolero, an R package I’ve built for getting a research project’s early, easy-to-get-wrong decisions right the first time — where the data lives, whether dependencies are tracked, whether a document and its extracted script can drift apart.

Rather than talk about the package in the abstract, we’ll spend most of the hour with it open in front of us, working through one small running example — bill and flipper measurements from the Palmer Penguins dataset — live:

1. init_project() and create_qmd(). Scaffolding a full project structure, renv, git, and a literate Quarto document in two calls, including the post-render hook that keeps that document and its extracted .R script permanently in sync.
2. detect_execution_context(). One function that lets the same analysis run correctly whether it’s stepped through interactively, rendered, or called unattended via Rscript — say, from a cron job or a cluster.
3. A split-apply-combine workflow. write_by_group(), run_by_group(), and save_output(), for anyone whose analysis needs to run per-subset rather than all at once — and the direct on-ramp to high-throughput computing when a laptop stops being enough.

Bring a laptop if you’d like to follow along rather than just watch — this is a “show and try” session, not a lecture.

How to prepare

Nothing is required beforehand, but a couple of minutes now will make the live-coding go smoother:

• Install toolero from CRAN: install.packages("toolero").
• If you’d rather install the development version, or just want to poke around the source first, the GitHub repository has the full README and function reference.
• Have R and RStudio (or your editor of choice) open and ready to go — we’ll be running code together, not just watching it run.

None of this is required reading. I’ll walk through everything live, but having toolero already installed means you can jump in the moment we get to the first demo instead of waiting on an install partway through.