Why is fill not working in my ggplot in my shiny app?

Question

I'm trying to replicate the following code into my shiny app:

df1 <- df %>% 
  group_by(xyz, inception_month) %>% 
  summarise(policies = sum(policies))

a <- ggplot(df1, aes(x=as.factor(inception_month), y = policies,  fill = xyz)) +
  geom_bar(position = "fill", stat = "identity") + 
  theme(axis.text.x = element_text(angle = 30, hjust = 1)) +
  xlab("Inception Month") +
  ylab("Policy Count")
print(a)

which gives me the following desired output: enter image description here

However I can't get the fill to display with the following dynamic code:

server <- function(input, output) {

df1 <- reactive({df %>% group_by(input$rating_factor, inception_month) %>% summarise(policies = sum(policies))})

  output$plot <- renderPlot({
    p <- ggplot(data = df1(), aes(x=as.factor(inception_month), y = policies, fill = input$rating_factor)) +
      geom_bar(position = "fill", stat = "identity") +
    theme(axis.text.x = element_text(angle = 30, hjust = 1)) +
      xlab("Inception Month") +
      ylab("Policy Count")

    print(p)
  })
}
shinyApp(ui, server)

df looks like:

df <- data.frame(
  xyz = c("A","A","B","B","B"),
  abc = c("X","X","Y","Y","Y"),
  inception_month = c("Feb 2020", "Nov 2019", "Feb 2020", "Dec 2019", "Feb 2020"),
  policies = c(1, 0, 1, 1, 1))

and the ui is:

library(shiny)
library(shinydashboard)
library(dplyr)
library(ggplot2)

ui <- dashboardPage(
  dashboardHeader(),
  dashboardSidebar(),
  dashboardBody(
    fluidRow(
    box(plotOutput("plot", height = 250)),

    box(
      title = "Select Rating Factor",
      selectInput("rating_factor", "Rating Factor", c("xyz", "abc"))
  )
  )
))

Hello, Antzlck. Welcome to StackOverflow. The example you posted isn't reproducible. We don't know what "df1" is made of. We have no access to "df". To improve your chances of getting a reply, you should either share the data in some way along with your code, or better, generate some fake data and make sure the column names in the fake data match the ones from your code. In R, the reprex package is great if you want to create a reproducible example. — Jxson99, Apr 23 '20 at 14:28

bretauv · Accepted Answer · 2020-04-23T17:03:49.133

input$rating_factor creates a character value, which is not what ggplot expects. You need to use !!sym(input$rating_factor) instead.

I agree with Mason Beau's comment, though. Next time, you should provide a reproducible example. See here or here to know how to make one.

Edit: As detailed here, you have to use !!sym() in all dplyr functions as well. Here's your corrected example:

library(shiny)
library(shinydashboard)
library(dplyr)
library(ggplot2)

df <- data.frame(
  xyz = c("A","A","B","B","B"),
  abc = c("X","X","Y","Y","Y"),
  inception_month = c("Feb 2020", "Nov 2019", "Feb 2020", "Dec 2019", "Feb 2020"),
  policies = c(1, 0, 1, 1, 1))

ui <- dashboardPage(
  dashboardHeader(),
  dashboardSidebar(),
  dashboardBody(
    fluidRow(
      box(plotOutput("plot", height = 250)),

      box(
        title = "Select Rating Factor",
        selectInput("rating_factor", "Rating Factor", c("xyz", "abc"))
      )
    )
  ))

server <- function(input, output) {

  df1 <- reactive({
    df %>% 
      group_by(!!sym(input$rating_factor), inception_month) %>% 
      summarise(policies = sum(policies))
  })

  output$plot <- renderPlot({
    p <- ggplot(data = df1(), aes(x=as.factor(inception_month), y = policies, fill = !!sym(input$rating_factor))) +
      geom_bar(position = "fill", stat = "identity") +
      theme(axis.text.x = element_text(angle = 30, hjust = 1)) +
      xlab("Inception Month") +
      ylab("Policy Count")

    print(p)
  })

}

shinyApp(ui, server)

Thanks. I've tried that, but I get the error `object 'xyz' not found` — Antzlck, Apr 23 '20 at 16:07
@Antzlck you need to make this change in all `dplyr` functions, like `group_by` in your example. I edited my answer — bretauv, Apr 23 '20 at 17:04

Why is fill not working in my ggplot in my shiny app?

1 Answers1