Modular, layout-as-code approach for customizable Shiny dashboards

on YouTube ↗18m

" summary

A Shiny dashboard is split into reusable page/tile/content modules so GDPR-sensitive bits are granted per-module, not per-app, using a layout-as-code JSON that stores per-user layouts in Postgres; module communication uses reactive exports/imports (R reactive values passed by reference) so graphs update when a search selection changes. Runtime drag‑and‑drop persists to the JSON in the DB, the app runs on Posit Connect, a few hundred users use it, four developers maintain an R6-based framework that makes adding modules one‑or‑two files and it works in production.

AI-written summary · about

" chapters (11)

1. 0:00Healthcare regulation dashboards
2. 1:06Healthcare system trade-offs
3. 3:16Privacy law and modular design
4. 5:04Spaghetti vs waffle code
5. 6:30Shiny modules for dashboards
6. 8:16Layout-as-code with JSON
7. 10:27Module communication via JSON
8. 13:00Runtime customization
9. 14:47Modular design benefits
10. 15:32Resources for Shiny dashboards
11. 16:15Q&A

" related talks

• 16mSmall Team Large Organization: Building Impactful Shiny Dashboards at NIHJon Nye · rstudio::conf(2022)
• 32mDash: Interactive Data Visualization Web Apps with no JavascriptDom Weldon · EuroPython 2019
• 14mR Shiny - From Conception to the CloudIvonne Carrillo Dominguez · rstudio::conf(2022)
• 23mEmpowering Decisions: Advanced Portfolio Analysis and Management through ShinyLovekumar Patel · posit::conf(2024)
• 35mBuild Simple and Scalable Apps with ShinyGordon Shotwell, Tracy Teal · PyData NYC 2023
• 14mR Shiny - From Conception to the CloudIvonne Carrillo Dominguez · rstudio::conf 2019
• 20mTowards the Next Generation of Shiny UICarson Sievert · posit::conf(2023)
• 09mPublishing Customer Facing Products with RStudio ConnectBenjamin Braun · rstudio::conf 2019