Make your code shiny

I have recently written a very basic beginner's guide to a simple shiny app. To be found here

The intention was that I had a set of results (differing by the value of one input variable) that I wanted to compare. After some tiresome rerunning the model and plotting results for comparisons I decided to make the whole process a bit quicker.

Thanks to shiny adjustment of the input values created results at an instant, made the whole interactive and more fun to present.

Sounds good? Ok, lets do it:

Some very basic introduction to build your shiny app.

Get started...

To start off install.packages("shiny"). The two main components of your first shiny app are the ui (user interface) and server (holds the server-side logic - the server function). Basically you take care of all your designs in the ui and add all your functionality in the server (user input ==> outputs). Now two options are available:
1-store files separately in one folder as ui.R and server.R
2-store them in one file (in this case make sure to add shinyApp(ui = ui,
server = server) in the end).

library(shiny)
ui <- fluidPage()
server <- function(input, output) {}
shinyApp(ui = ui, server = server)

I think shiny appears to be incredibly easy in creating webapps with lots of functionality and great looks, from what I have seen so far elsewhere (python, java). Though I have not tried out a lot in these languages and am far from confident in building python or java webapps.

To create a nice user interface shiny offers some easy to use (HTML derived) commands:
h1() add a first header - use h2, h3, h4, h5, h6 to add more headers,
strong() for bold text,
em() for italicized text,
br() to add a line break,
a() to add a hyperlink,
img() to add an image,
code() to display,
hr() for a horizontal rule.

Have a look here for more options.

Layout

There are several layout options:

1- Default layout...

...is to have a sidebar for inputs and a major panel for outputs (sidebarpanel(input), mainbarpanel(output) within the ui)

library(shiny)
library(ggplot2)
ui <- fluidPage(
h1("Some iris data visualised"),br(),

  sidebarLayout(position="left",

     sidebarPanel(
        sliderInput("bin", "Binwidth:",min = 0.1, max = 1, value = 1)
      ),

     mainPanel(plotOutput("IrisPlot"))
  )
)

server <- function(input, output) {
  output$IrisPlot <- renderPlot({
     ggplot(iris)+geom_histogram(aes(Sepal.Length, fill=Species),
          binwidth=input$bin, colour="white")+facet_wrap(~Species)
  })
}

shinyApp(ui = ui, server = server)

2- Grid layout...

...uses fluidrow(), columns() and wellpanel(). The page is separated into 12 columns, so make sure to add up to these. The wellpanel() creates a panel with grey background. To create the same image as above use:

library(shiny)
library(ggplot2)
ui <- fluidPage(
h1("Some iris data visualised"),br(),

  fluidRow(
     column(4,
       wellPanel(
         sliderInput("bin", "Binwidth:",min = 0.1, max = 1, value = 1)
       )
     ),

     column(8,
       plotOutput("IrisPlot")
      )
  )
)

server <- function(input, output) {
 output$IrisPlot <- renderPlot({
   ggplot(iris)+geom_histogram(aes(Sepal.Length, fill=Species),
      binwidth=input$bin, colour="white")+facet_wrap(~Species)
 })
}

Please have a look at this page to get more information on the Grid layout.

3- Layered layout...

...with tabsetPanel() and navlistPanel()

library(shiny)
library(ggplot2)
ui <- fluidPage(
  h1("Some iris data visualised"),br(),
  fluidRow(
   column(4,
      wellPanel(
        sliderInput("bin", "Binwidth:",
          min = 0.1, max = 1, value = 1)
    )),

   column(8,
     tabsetPanel(
       tabPanel("Plot",plotOutput("IrisPlot")),
       tabPanel("Summary",verbatimTextOutput("IrisSummary")),
       tabPanel("Table",tableOutput("IrisTable"))
    ))
 )
)
server <- function(input, output) {
 output$IrisPlot <- renderPlot({
     ggplot(iris)+geom_histogram(aes(Sepal.Length,         fill=Species),binwidth=input$bin, colour="white")+facet_wrap(~Species)
  })
 output$IrisSummary <- renderPrint({summary(iris)})
 output$IrisTable <-renderTable(iris)
}
shinyApp(ui = ui, server = server)

If you have more than a few tabs the navlistPanel may come in handy, adding sidebars instead of tabs to your shiny app. the code is the same as above, just replace tabsetPanel with navlistPanel.

4- Layout with several pages...

...navbarPage() will do the trick. "Just" replace your fluidPage() with navbarPage(), well..., I admit that is not all. You will have to restructure a bit, as shown below:

library(shiny)
library(ggplot2)
ui <-navbarPage("Iris data: ",
 tabPanel("Plot",
    sidebarLayout(position="left",
      sidebarPanel(
        sliderInput("bin", "Binwidth:",min = 0.1, max = 1, value = 1)
      ),
      mainPanel(
         plotOutput("IrisPlot")
      ))
 ),
 tabPanel("Summary",
    verbatimTextOutput("IrisSummary")),
 tabPanel("Table",
    tableOutput("IrisTable"))
 )

server <- function(input, output) {
     output$IrisPlot <- renderPlot({
        ggplot(iris)+geom_histogram(aes(Sepal.Length, fill=Species),
            binwidth=input$bin, colour="white")+facet_wrap(~Species)
     })
 output$IrisSummary <- renderPrint({summary(iris)})
 output$IrisTable <-renderTable(iris)
}

shinyApp(ui = ui, server = server)

Shiny done

Feel free to add more widgets (please have a look here for the vast amount of choices) and off you go with your first basic shiny App

code below...

For more information have a look at this really nice tutorial (very helpful).

library(shiny)
library(ggplot2)
ui <- fluidPage(
 h1("Some iris data visualised"),br(),
  fluidRow(
    column(4,
        wellPanel(
          sliderInput("bin", "Binwidth:",
            min = 0.1, max = 1, value = 1)
    )),
    column(8,
      tabsetPanel(
        tabPanel("Plot",plotOutput("IrisPlot")),
        tabPanel("Summary",verbatimTextOutput("IrisSummary")),
        tabPanel("Table",tableOutput("IrisTable"))
    ))
 ),
 fluidRow(
    column(4,
      checkboxGroupInput("Species", label = "Select Species",
        choices = list("Iris setosa" = "setosa", 
          "iris versicolor" = "versicolor",
          "iris verginica" = "virginica"),
          selected = 1)),
    column(8,plotOutput("IrisPlot2"))
 )
)

   server <- function(input, output) {
     output$IrisPlot <- renderPlot({
        ggplot(iris)+geom_histogram(aes(Sepal.Length,     fill=Species),binwidth=input$bin, colour="white")+facet_wrap(~Species)
     })
 output$IrisSummary <- renderPrint({summary(iris)})
 output$IrisTable <-renderTable(iris)

output$IrisPlot2 <- renderPlot({
   ggplot(iris[iris$Species==input$Species,])+geom_point(aes(Sepal.Length,     Sepal.Width,color=Species))
 })
}

shinyApp(ui = ui, server = server)