You just attended my talk « Shiny: from prototype to production » at Rencontres R 2026? Welcome, you’re in the right place.
This page gathers everything: the slides, and for each of the 6 steps, the resource to dig into the topic at your own pace.
📥 The slides
🎥 The talk video
The replay will be added here as soon as it’s available. To be notified, subscribe to the newsletter (right at the bottom of this page).
The 6 steps, and where to dig deeper
The talk gave you the map. Here, for each step, is the article that goes into detail.
1. Project structure
Get out of the single file. We separate responsibilities: the business logic (model / service) on one side, the interface (the modules) on the other, plus tests and well-organized assets.
➡️ How to Structure Your Shiny App Like a Pro
2. Reproducibility
“It works on my machine” isn’t a deployment strategy. We pin our dependencies with renv, our secrets in a .Renviron, and our R version.
➡️ How to use renv to make your Shiny apps reproducible?
3. Code quality
Code that works isn’t enough, you need code that lasts. We set ourselves code standards, automate their enforcement with lintr and styler, then tackle the substance: clear naming, factoring (DRY), short functions, clean error handling, and first tests.
➡️ How to Ensure Code Quality in Your Shiny Applications
4. Software design
Software that lasts is software that was thought through. We design the screens and the journeys before coding, and we treat the UI as real web dev (bslib + Sass).
➡️ How to Improve the Design of Your Shiny Application
5. Deployment
No need for AWS or an over-engineered setup: a VM for a few euros a month, a Docker image, a CI that deploys on every push, and a load test before opening to the public.
➡️ How to Deploy a Shiny App on a VM with Hetzner
6. Security
Don’t write your authentication yourself. A service like Auth0 integrates cleanly around your app, with HTTPS and a hardened server.
➡️ Your Definitive Guide to Seamless Authentication with Auth0 in Shiny Apps
What now?
You don’t have to do everything at once. Pick one step, read the resource, apply it. Then the next one.
And if you have a Shiny application to take to production and want to discuss it, get in touch, or come say hi at the Data Champ’ booth during the conference.
Comments