ggplot and shiny: ggplot doesn't know how to deal with reactive: yes it is NOT a repeat

Question

I am trying to use ggplot2 with a reactive object in shiny. So, I understand that reactive gives a reactive object that needs to be changed to be a value, but I tried to assign it to a different object both in the reactive part and in the renderPlot part and got no results...

So, if I don't assign it to a different variable, I get the message:

ggplot2 doesn't know how to deal with data of class 
reactiveExpr/reactive

I know there are at least 2 questions (here and here) that talk about this problem. I did do my homework and try to use those answers and it didn't work. As far as I can tell, those answers suggest to assign the reactive object to another variable, which I did, and got the message:

Error: object is not a matrix

The answers delineated there do not really explain how to fix the problem in general, they just provide the code to fix that particular example (one answer reads " I think this particular problem should be solved with the following code:" but doesn't point out what part of the code solves the issue).

Can you please explain, how to fix this in this example in the more general case?

here is a Minimal example:

server.R

library(shiny)
shinyServer(function(input, output) {
  library(ggplot2)
  library(lme4)
  model1 <- lmList((conc) ~ time | Subject, data = Indometh)
  newpoints <- reactive({data.frame("Subject" = c(1, 4, 2, 5, 6, 3),
                          "conc" = predict(model1, newdata = data.frame(time=input$sliderTime)),
                          "time" = rep(input$sliderTime, 6))
  })

  output$myPlot <- renderPlot({
    newpoints2 <- newpoints()
    g <- ggplot(Indometh, aes(time, (conc), color= Subject)) + geom_point()  +
      stat_smooth(method="lm",fullrange=TRUE, fill=NA)
    newg <- g + geom_point(data = newpoints2, mapping =
                             aes(x = time, y = (conc), color= factor(Subject)))
    print(newg)

  })

})

and the ui.R file:

library(shiny)
shinyUI(fluidPage(

  # Application title
  titlePanel("Concentration of Indomethacin"),
  sidebarLayout(
    sidebarPanel(
      sliderInput("slidertime",
                  "Time:",
                  min = 8,
                  max = 15,
                  value = 8)),

    mainPanel(
       plotOutput("myPlot")
    )
  )
))

Alejandro Andrade · Accepted Answer · 2018-03-23T21:35:40.690

The problem with your example is that you are calling the input with the wrong name it is not sliderTime it is slidertime i change the server to this and it works perfect.

How I fix it? just use browser() inside the reactive expression and found that input$sliderTime was NULL so I check the name.

library(shiny)
shinyServer(function(input, output) {
  library(ggplot2)
  library(lme4)
  model1 <- lmList((conc) ~ time | Subject, data = Indometh)
  newpoints <- reactive({
    data.frame("Subject" = c(1, 4, 2, 5, 6, 3),
               "conc" = predict(model1, newdata = data.frame(time=input$slidertime)),
               "time" = rep(input$slidertime, 6))
  })

  output$myPlot <- renderPlot({

    newpoints2 <- newpoints()
    g <- ggplot(Indometh, aes(time, (conc), color= Subject)) + geom_point()  +
      stat_smooth(method="lm",fullrange=TRUE, fill=NA)
    newg <- g + geom_point(data = newpoints2, mapping =
                             aes(x = time, y = (conc), color= factor(Subject)))
    print(newg)

  })

})

Interesting solution, I had no idea you could use `browser` inside a shiny session — divibisan, Mar 08 '19 at 23:07
Yeah its like the only way I have found to debug a shiny app, specially if everything is reactive. — Alejandro Andrade, Mar 09 '19 at 01:48

ggplot and shiny: ggplot doesn't know how to deal with reactive: yes it is NOT a repeat

1 Answers1