Creating Shiny apps with R

7. Allow users to upload and download the data

i. Allow users to upload and download data from an app

• It's simple
• downloadButton() in ui.R
• downloadHandler() in server.R
• ​paste0(Sys.Date(), "_bechdel-ratings.csv")​

ii. Download data from DT tables

• Most flexible soln to download data from table is to used DT package
• Extensions argument
• Dom argument - Bftip

iii. Allow users to upload data to an app

Will be useful for map generation

iv. Uploading files into Shiny App

• Use fileInput("uploaded_file") to upload files in ui.R
• In server.R, use input$uploaded_file$datapath to access the uploaded file
• Uploaded files will be created in a temporary directory, and will be deleted when the app exits
  • If you want to keep the file permanently, it should be explicitly stated in the code
• Need to create a reactive variable in the server.R to store the uploaded file
  • uploaded_file <- reactive({read.csv(input$uploaded_file$datapath)})
  • input$uploaded_file$datapath is a reactive expression, so it needs to be wrapped in reactive()

v. Using rhandsontable to collect data

• rhansontable is a htmlwidget package
• rhansontable library should be loaded in both ui.R and server.R

8. Problem solving in Shiny apps

====================================

i. Problem solving in Shiny apps 101

• Misspelling the variables between ui.R and server.R

• Deploying shiny apps on making available to colleagues and customers

• Shiny app that works on your local machine might not necessarily work when deployed

• Three skills you need

  1. Debugging R code
  2. Debugging web content (HTML, CSS, JavaScript)
  3. Debuggin deploying issues

• Debugging steps R code

  • Use print() to send variables/data to the console
  • Use shinyjs::runCodeUI() and shinyjs::runcodeServer() to interactively explore R variables in your Shiny app
  • Use shiny::reactLog() to explore the issues with reactivity

• Debugging web components

  • Use web browser's inspector to investigate the issues and change CSS/HTML attributes dynamically
  • shinyjs::showLog() to show the log of JavaScript events

• Appropriate packages are loaded in ui.R and server.R

• All file paths must be relative to Shiny app's root directory

• Inlude all code in the server.R or use source("data-processing.R", local = TRUE) to source external files

• Check issues is not due to difference in time zones in the date function

• When using APIs, ensure tokens are sent to the server and are not locked in the local machine

====================================

ii. Printing to the R console in Shiny apps

• We cannot access the value of input variables in outside reactive expressions
• You can use print function to within reactive expressions like render plot, event reactive, observe event and its cousins
  • The examples of reactive expressions are renderPlot(), eventReactive(), observeEvent(), observe(), reactive(), reactiveValues()

ii. Debugging Shiny apps with shinyjs::runcodeUI

• Load shinyjs library in both ui.R and server.R
• Add folowing block of code to ui.R inside fluidPage()
  • useShinyjs()
  • runcodeUI(type = "textarea")
• Add runCodeServer() to server.R inside shinyServer() function

iii. Using reactlog to debug Shiny apps

• reactlog is a package that helps to debug reactive components of the Shiny app
• options(shiny.reactlog = TRUE) in R console to enable reactlog
• Ctrl + F3 to open reactlog or Cmd + F3 in Mac

====================================

9. Making Shniy apps beautiful

i. Applying custom CSS to Shiny apps

• ==CSS: Cascading Style Sheets== - all web contents are styled using CSS
• CSS can be applied "inline" to specific Shiny elements
• Style and entire app using theme argument of fluidPage()/navbarPage()
• Always save CSS and JavaScript files in www subfolder folder
• Themes are all changed within the ui.R file
• Problem: Pull down menu is below the zoom (+/-) button
• Solution: Add the selectInput() within the div() function and add style = "postion:relative;z-index: 1000;" to the div() function

ii. Inserting images to Shiny apps

• Different ways to insert images in Shiny app
• img(src = "images/plot.png") in ui.R
• RenderImage() in server.R and imageOutput() in ui.R
• HTML tags in ui.R - <img> </img> tags

iii. Show loading spinners in Shiny apps

• Good idea to use spinners when a function takes couple of seconds to load
• Very nice function; could be used in ==mapping Shiny app==
• withLoader() function in shinycustomloader package in ui.R file

====================================

10. Shiny app deployment

i. Where can you deploy shiny apps?

• Value in Shiny apps is in sharing them with others
• Two options for hosting
  • Self-hosting
  • Cloud hosting
• Self-hosting is a bit of a misnomer as it usually means running server applications on a virtual private server (VPS) like Amazon Web Services (AWS), Azure, or Digital Ocean
• Cloud hosting - sinyapps.io, RStudio Connect, Shiny Server Pro
• Shiny apps if used commercially might require a self hosted architecture due to data provisioning requirements
• shinyapps.io is not currently GDPR compliant and unlikely to be in the near future
  • GDPR - General Data Protection Regulation

ii. Deploying Shiny apps to shinyapps.io

• Copy the secret key from the shinyapps.io account to RStudio while publishing the app
• The more complicated your app is, the more time it takes to deploy
• Here is your deployed test app: https://himal-shrestha.shinyapps.io/testapp/

iii. Managing Shiny apps with RStudio Connect

iv. Programmatically deploying Shiny apps

Jun 28

Made shiny app to accept inputs and then plot the map of the country