Why the histogram doesn't work - shiny library in R

Question

Dataset: https://www.kaggle.com/mirichoi0218/insurance

Hi, I have a problem with shiny library. Wanted to create app that will show the Charges depending on the age, region and children. Unfortunately, the histogram doesn't work.

ui <- fluidPage(
  titlePanel("Insurance charges"),
  sidebarLayout(
    sidebarPanel(
      sliderInput("ageInput", "Age", 18, 64, 41),
      radioButtons("childrenInput", "Liczba dzieci",
                   choices = c("0", "1", "2", "3",'4','5'),
                   selected = "0"),
      selectInput("regionInput", "Region",
                  choices = c("NE", "NW", "SE","SW"))
    ),
    mainPanel(
      plotOutput("coolplot"),
      br(), br(),
      tableOutput("results")
    )
  )
)

server <- function(input, output) {
  output$coolplot <- renderPlot({
    filtered <-
      insurance %>%
      filter(age >= input$ageInput[1],
             age <= input$ageInput[2],
             children == input$childrenInput,
             region == input$regionInput
      )
    ggplot(filtered, aes(insurance$charges)) +
      geom_histogram()
  })
}

shinyApp(ui = ui, server = server)

Welcome to stackoverflow.com. Please read how to produce an MCVE. Specify the problem and describe the desired result and your attempts to reach it. — asd-tm, Jan 11 '22 at 17:17
`ggplot(filtered, aes(insurance$charges))` is wrong and likely comes with clear errors (that you've not included in your question). Perhaps you meant `ggplot(filtered, aes(charges))`? What you've done is akin to `ggplot(mtcars, aes(1:3))`. The aesthetic you've assigned is very likely not the same length as the data that `ggplot` intends to use, so `Aesthetics must be either length 1 or the same as the data (32): x` should be a clear error. If you are not getting *that* error, then please don't keep it a secret from us. — r2evans, Jan 11 '22 at 17:38

score 2 · Accepted Answer · answered Jan 11 '22 at 17:40

The major issue with your code is that

you map insurance$charges on x in ggplot(). That's why you always get a histogram for the total dataset. Simply use aes(charges) to use the column from your filtered dataset.
Not sure about that. Maybe your recoded your data, but the regions in the Kaggle dataset have labels "northeast", "northwest", .... To account for that you could use a named vector for the choices in the selectInput
Finally, you have to use a vector of length 2 for the value of the sliderInput to get a range slider.

And as an example on how to provide a minimal reproducible example I used a random fake dataset for the reprex.

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

ui <- fluidPage(
  titlePanel("Insurance charges"),
  sidebarLayout(
    sidebarPanel(
      sliderInput("ageInput", "Age", min = 18, max = 64, value = c(18, 41)),
      radioButtons("childrenInput", "Liczba dzieci",
        choices = c("0", "1", "2", "3", "4", "5"),
        selected = "0"
      ),
      selectInput("regionInput", "Region",
        choices = c("NE" = "northeast", "NW" = "northwest", "SE" = "southeast", "SW" = "southwest")
      )
    ),
    mainPanel(
      plotOutput("coolplot"),
      br(), br(),
      tableOutput("results")
    )
  )
)

server <- function(input, output) {
  output$coolplot <- renderPlot({
    filtered <- insurance %>%
      filter(
        age >= input$ageInput[1],
        age <= input$ageInput[2],
        children == input$childrenInput,
        region == input$regionInput
      )

    ggplot(filtered, aes(charges)) +
      geom_histogram()
  })
}

shinyApp(ui = ui, server = server)

DATA

set.seed(123)

insurance <- data.frame(
  age = sample(18:64, 1000, replace = TRUE),
  children = sample(0:5, 1000, replace = TRUE),
  region = sample(c("southwest", "southeast", "northwest", "northeast"), 1000, replace = TRUE),
  charges = runif(1000, 3000, 40000)
)

Why the histogram doesn't work - shiny library in R

1 Answers1