This workshop is designed as an overview of some of the tools for writing Shiny apps. Shiny is an R package for writing web applications around your code and data. If you've never heard of Shiny before, have a quick look at this excellent gallery of example web apps that have been curated by Rstudio.
Here are the packages that we'll use in this workshop:
# The absolute minimum
install.packages("shiny")
install.packages("ggplot")
install.packages("plotly")
install.packages("dplyr")
# Package for profiling code
install.packages("profvis")
# Packages for making your application look different
install.packages("shinydashboard")
install.packages("bslib")If you are familiar with writing R code, then you probably won't have much trouble converting it to work with user selectable inputs. In fact, that's how this workshop starts! We'll look at an R notebook that makes cool example visualizations with a toy dataset. Then we'll look at what it takes to convert that R notebook into a functioning Shiny app.
At first we focused mostly on just converting our R code into something that works with Shiny in the simplest ways possible. We got something that works, but there may be more efficient ways to structure our code to run better in the Shiny context. That's exactly the kind of thing we can test by profiling our code with profvis.
Next, we'll start thinking about reactivity, which is the underlying logic that controls how Shiny works, when and how it updates the outputs we've specified.
After reactivity, we'll go over customizing the look of your Shiny app with packages like shinydashboard and bslib. If there's time, we'll go through writing your own CSS to customize the look of your app.
Finally, we'll cover a few options for deploying your Shiny app to the web.
We will be using Health Canada's nutrient dataset to look at nutrient values for a variety of ingredients. This workshop is based off of yihanwu's excellent Shiny app.
I picked this dataset because the amount of stuff to explore in it is immense. There are over 5000 food items to select, and for each of them we can look at the associated nutrient information. There are many ways you could choose to visualize the nutrition data. For instance, we could look at the raw nutritional value vs the percent of the daily recommended value. We could plot the information or put it in a table. We could even combine food items and estimate their total nutritional content.
I suspect that's all we'll have time to get to, but that won't stop me from thinking about some more advanced topics to potentially cover in the future! Here's what I'm considering:
- Dynamic UIs
- Conditional UIs
- Performant code for heavy computation/simultaneous users
- Structuring Shiny deployments with Docker
If you'd like to suggest topics to integrate into this workshop, you can email me or file it as an issue with this repository.