CODEWITHSUNDEEP
rggplot2shinyshiny-reactivity
Diego Merlano Porto
Diego Merlano Porto

Reputation: 15

Why do I have to use aes_string() with ggplot in shiny?

I´m new working with shiny and I´m trying to make a ggplot. According to what I have read, I have to use aes_string instead of aes(), but I am still trying to figure out why.

My code:

output$scatterplot <- renderPlot({
    ggplot(data = movies, aes_string(x = input$x, y = input$y,
                                     color = input$z)) +
      geom_point(alpha = input$alpha)
  })

Upvotes: 1

Views: 1179

Answers (1)

Ronak Shah
Ronak Shah

Reputation: 388982

Let us take mtcars data as an example.

Normally you call ggplot2 as :

library(ggplot2)
ggplot(data = mtcars, aes(x = mpg, y = disp,color = am)) + geom_point()

enter image description here

Notice that you pass mpg, disp and am as bare column names and not as string. When you take user input in shiny you get those as string. So if you don't use aes_string what you are calling is :

ggplot(data = mtcars, aes(x = "mpg", y = "disp",color = "am")) + geom_point()

enter image description here

which is not what you want. Hence, by using aes_string you "tell" R/Shiny that I am passing column names as string but treat them as columns from the dataframe and not as string.

Also note that aes_string has been soft-deprecated and it is recommended to use tidy evaluation using .data as @Phil mentioned.

ggplot(data = mtcars, 
       aes(x = .data[["mpg"]], y = .data[["disp"]],color = .data[["am"]])) + 
  geom_point()

Other options include :

  • sym with !! :
ggplot(data = mtcars, 
     aes(x = !!sym("mpg"), y = !!sym("disp"),color = !!sym("am"))) + 
     geom_point()
  • get :
ggplot(data = mtcars, 
       aes(x = get("mpg"), y = get("disp"),color = get("am"))) + 
  geom_point()

Upvotes: 8

Related Questions