Deploying End-To-End Data Science with Shiny, Plumber, and Pins
summary
Builds a Shiny app that predicts bike counts using Capital Bikeshare API, trains an XGBoost model daily, and schedules data ingest every 20 minutes; uses R Markdown for reproducible pipelines, serves the model via a Plumber API, and stores station metadata (~600 rows) and model objects with Pins (good for items < ~1 GB). Deploys on RStudio Connect for auth, scheduling, and serving or uses any R-enabled server with proxies/firewalls for security. Shiny, Plumber, Pins RStudio Connect
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:24The bike share app
- 1:40Deployment goals
- 3:10The need for a better solution
- 5:35Introducing the pins package
- 8:25Best practices for pinning
- 10:34Pins in action
- 14:15Securing the application
- 16:27Key takeaways
- 17:07Q&A session
related talks
23mEmpowering Decisions: Advanced Portfolio Analysis and Management through Shiny
15mR Markdown + RStudio Connect + R Shiny
14mR Shiny - From Conception to the Cloud
15mA Fiery approach to R web services
35mBuild Simple and Scalable Apps with Shiny
01h 44mShinylive - Running R and Python Dashboards without a Server
15mA Robust Framework for Automated Shiny App Testing
15mI made an entire e-commerce platform on Shiny