在R Shiny中实现三元柱状图切换为指定二元柱状图

问题背景 #

我有包含国家、市中心一居室房租（x48）、郊区一居室房租（x49）、月平均薪资（x54）的数据集，原本用ggplot生成了三元柱状图，但现在需要固定保留月平均薪资的柱子，同时通过Shiny输入控件选择显示x48或x49其中一个，实现二元柱状图的动态切换。

数据集结构如下：

data <- structure(list(country = c("Andorra", "Austria", "Belgium", "Denmark", "Finland", "France", "Germany", "Greece"), 
                       x48 = c(811.73, 710.0385, 765.8359, 961.434, 701.9176, 716.7308, 701.6187, 468.1942), 
                       x49 = c(655.2433, 607.8058, 628.3675, 692.426, 575.4287, 587.7711, 544.2052, 370.2103), 
                       x54 = c(2801.597, 2224.757, 2320.216, 3130.023, 2582.625, 2399.011, 2621.368, 804.5603)), 
                  class = "data.frame", row.names = c(NA, -8L))

解决方案 #

核心思路是根据Shiny输入筛选需要保留的两个变量（固定x54+选中的x48/x49），再进行数据重塑和绘图。以下是完整的可运行Shiny代码：

library(shiny)
library(tidyverse)

# 加载数据集
data <- structure(list(country = c("Andorra", "Austria", "Belgium", "Denmark", "Finland", "France", "Germany", "Greece"), 
                       x48 = c(811.73, 710.0385, 765.8359, 961.434, 701.9176, 716.7308, 701.6187, 468.1942), 
                       x49 = c(655.2433, 607.8058, 628.3675, 692.426, 575.4287, 587.7711, 544.2052, 370.2103), 
                       x54 = c(2801.597, 2224.757, 2320.216, 3130.023, 2582.625, 2399.011, 2621.368, 804.5603)), 
                  class = "data.frame", row.names = c(NA, -8L))

# 预定义变量标签与颜色映射
var_labels <- c(
  x48 = "1 room in the center",
  x49 = "1 room outside the center",
  x54 = "Average Monthly Salary"
)
var_colors <- c(
  x48 = "#108CDF",
  x49 = "#0DD016",
  x54 = "#DA560A"
)

# UI界面
ui <- fluidPage(
  selectInput(
    inputId = "selected_rent",
    label = "选择要显示的房租变量",
    choices = c("市中心一居室" = "x48", "郊区一居室" = "x49"),
    selected = "x48"
  ),
  plotOutput("bar_plot")
)

# 服务器逻辑
server <- function(input, output) {
  output$bar_plot <- renderPlot({
    # 筛选需要的变量：固定x54 + 用户选中的房租变量
    selected_vars <- c(input$selected_rent, "x54")
    
    data %>%
      select(country, all_of(selected_vars)) %>%
      pivot_longer(-country, names_to = "metric", values_to = "value") %>%
      ggplot(aes(x = country, y = value, fill = metric)) +
      geom_col(position = "dodge") +
      scale_fill_manual(
        name = "Metric",
        labels = var_labels[selected_vars],
        values = var_colors[selected_vars]
      ) +
      theme_light() +
      theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1)) +
      xlab("Country") +
      ylab("USD")
  })
}

shinyApp(ui, server)

关键修改说明 #

• 变量筛选：通过select(country, all_of(selected_vars))只保留需要的列，避免多余数据进入绘图流程。
• 动态匹配标签与颜色：提前定义映射表，根据选中的变量动态提取对应内容，保证图例和颜色始终匹配当前显示的变量。
• 实时响应输入：通过input$selected_rent获取用户选择，实时更新绘图数据，实现二元柱状图的切换效果。

内容的提问来源于stack exchange，提问作者walter