Is it possible to initialize brush in ggplot in a shiny app?

Question

I have a shiny app and I want a ggplot with a brush at the start, so the user does not need to select certain area of interest every time the app starts. Later of course user can select different area. Here is an example to start with:

library(shiny)
library(ggplot2)

runApp(shinyApp(
  ui = fluidPage(plotOutput('plotA', brush = brushOpts(id = 'plotA_brush')),
                 plotOutput('plotZ')),
  server = function(input, output, session) {
    pollData <- reactivePoll(60 * 1000, session,
                             checkFunc = function(){ Sys.time() },
                             valueFunc = function(){ data.frame(x = 1:100, y = cumsum(rnorm(100)))})
    output$plotA <- renderPlot({
      dt <- pollData()
      ggplot(dt, aes(x, y)) + geom_line()
    })
    ranges <- reactiveValues(x = NULL, y = NULL)
    observe({
      brush <- input$plotA_brush
      if(!is.null(brush)) {
        ranges$x <- c(brush$xmin, brush$xmax)
        ranges$y <- c(brush$ymin, brush$ymax)
      } else {
        ranges$x <- NULL
        ranges$y <- NULL
      }
    })
    output$plotZ <- renderPlot({
      dt <- pollData()
      ggplot(dt, aes(x, y)) + geom_line() + coord_cartesian(xlim = ranges$x, ylim = ranges$y)
    })
  }
))

Answer 1

Yes, it is possible.

In the code below I've added just a few lines. First, I've added set.seed(42) so that the graphics are reproducible. Second, there is a dput(brush) that has been commented out. This was helpful for identifying the initial brush that I wanted to have. Lastly, The in the observe environment controlling the ranges I have added a if else set to use the default brush is it is a NULL value from the input$plotA_brush object.

library(shiny)
library(ggplot2)
set.seed(42)

runApp(shinyApp(
  ui = fluidPage(plotOutput('plotA', brush = brushOpts(id = 'plotA_brush')),
                 plotOutput('plotZ')),
  server = function(input, output, session) {


    pollData <- reactivePoll(60 * 1000, session,
                             checkFunc = function(){ Sys.time() },
                             valueFunc = function(){ data.frame(x = 1:100, y = cumsum(rnorm(100)))})
    output$plotA <- renderPlot({
      dt <- pollData()
      ggplot(dt, aes(x, y)) + geom_line()
    })
    ranges <- reactiveValues(x = NULL, y = NULL)
    observe({
      if (is.null(input$plotA_brush)) {
        brush <- structure(list(xmin = 14.313925002001, xmax = 39.942241912585, ymin = 1.1077251080591, ymax = 5.5028180250535, mapping = structure(list( x = "x", y = "y"), .Names = c("x", "y")), domain = structure(list( left = -3.95, right = 104.95, bottom = -4.07771077213569, top = 9.69030145758825), .Names = c("left", "right", "bottom", "top")), range = structure(list(left = 32.3904099935947, right = 674.020527857828, bottom = 368.859578048224, top = 5.47945189149413), .Names = c("left", "right", "bottom", "top")), log = structure(list(x = NULL, y = NULL), .Names = c("x", "y")), direction = "xy", brushId = "plotA_brush", outputId = "plotA"), .Names = c("xmin", "xmax", "ymin", "ymax", "mapping", "domain", "range", "log", "direction", "brushId", "outputId"))
      } else {
        brush <- input$plotA_brush
      }
      # dput(brush)  # Useful for finding the initial brush
      if(!is.null(brush)) {
        ranges$x <- c(brush$xmin, brush$xmax)
        ranges$y <- c(brush$ymin, brush$ymax)
      } else {
        ranges$x <- NULL
        ranges$y <- NULL
      }
    })
    output$plotZ <- renderPlot({
      dt <- pollData()
      ggplot(dt, aes(x, y)) + geom_line() + coord_cartesian(xlim = ranges$x, ylim = ranges$y)
    })
  }
))

The initial starting page looks like this: