graph not displaying

[nnini]

Trying to sketch a stacked barplot with shiny app in R and would like some assistance. There are two selectinputs, one for the x axis which are agegp, tobgp and alcgp and the distribution variable which are alcgp and tobgp. I just cant quite figure out why its not displaying.

Any help would be greatly appreciated!

I have also provided a sample data set for esoph

esoph <- data.frame(
  agegp = c("18", "35", "65"),
  alcgp = c("Never", "Current", "Former","Current", "Former","Current"),
  tobgp = c("Never", "Current", "Former", "Never", "Current","Former"),
  cases = c(50, 75, 30, 40, 60, 25)
)

  ui <- fluidPage(
  titlePanel("Oesophagus Cancer Risk Dashboard"),
  tabsetPanel(

    # Stacked Bar Graph
    tabPanel("Stacked Bar Graph",
             fileInput("upload_graph_data", "Upload CSV file"),
             downloadButton("download_graph", "Export Graph (PNG)"),
             hr(),
             selectInput("x_variable", "X - Variable:",
                         choices = c("Age", "Tobacco Consumption", "Alcohol Consumption"),
                         selected = "Age"
             ),
             selectInput("distribution_variable", "Distribution Variable:",
                         choices = c("Tobacco Consumption", "Alcohol Consumption"),
                         selected = "Tobacco Consumption"
             ),
             plotOutput("stacked_bar_graph")
    )
    )
)
server <- function(input, output, session) {
  

  # Stacked Bar Graph

  filtered_data <- reactive({
    esoph_data <- esoph
    # Map X-Variable to appropriate column
    x_column <- switch(input$x_variable,
                       "Age" = "esoph$agegp",
                       "Tobacco Consumption" = "esoph$tobgp",
                       "Alcohol Consumption" = "esoph$alcgp"
    )
    # Filter by selected X-Variable
    esoph_data <- esoph_data[, c(x_column, input$distribution_variable)]
    # Filter by age group
    age_group_filter <- input$age_group
    if (age_group_filter[1] > 0) {
      esoph_data <- esoph_data %>%
        filter(agegp >= age_group_filter[1])
    }
    if (age_group_filter[2] < 75) {
      esoph_data <- esoph_data %>%
        filter(agegp <= age_group_filter[2])
    }
    return(esoph_data)
  })
  
  # Rendered  stacked bar graph
  output$stacked_bar_graph <- renderPlot({
    esoph_data <- filtered_data()
    esoph_data %>%
      group_by(.data[[x_column]], .data[[input$distribution_variable]]) %>%
      summarize(total_cases = sum(ncases)) %>%
      group_by(.data[[x_column]]) %>%
      mutate(percentage = 100 * total_cases / sum(total_cases)) %>%
      filter(!is.na(percentage) & percentage != 0) %>%
      ggplot(., aes_string(x = x_column, y = "percentage", fill = input$distribution_variable)) +
      geom_col(stat = "identity", position = "fill") +
      theme_minimal() +
      geom_text(aes(label = paste(format(percentage, digits = 1), "%")), size = 4, position = "fill", hjust = 0.5, vjust = 1.1) +
      scale_y_continuous(labels = scales::percent_format()) +
      labs(
        title = paste("Stacked Bar Chart of Case Distribution of", input$distribution_variable, "by", input$x_variable),
        subtitle = "Data Source: `esoph`",
        x = input$x_variable,
        y = "% of Cases",
        fill = input$distribution_variable
      )
  })
}

shinyApp(ui, server)

[nirgrahamuk]

@nnini

esoph_data <- esoph_data[, c(x_column, input$distribution_variable)]

on one side with x_column you have things like esoph$agegp which is not an appropriate column name to take from the esoph dataset, similarly input$distribution_variable can equal things like Tobacco Consumption which is also not a column name

one step in the direction, is to build in your own validation.
like this

 filtered_data <- reactive({
    esoph_data <- esoph
    # Map X-Variable to appropriate column
    x_column <- switch(input$x_variable,
                       "Age" = "esoph$agegp",
                       "Tobacco Consumption" = "esoph$tobgp",
                       "Alcohol Consumption" = "esoph$alcgp"
    )
    # Filter by selected X-Variable
    shiny::validate(
      shiny::need(x_column %in% names(esoph_data),
                  paste(x_column, "not in dataset")),
      shiny::need(input$distribution_variable %in% names(esoph_data),
                  paste(input$distribution_variable, "not in dataset"))
    )
    esoph_data <- esoph_data[, c(x_column, input$distribution_variable)]

then as a user when you run, you can get an easy to understand message about whats wrong. the solution is to have names you expect to access to match selectable columns in the data.