如何在ggplot翻转坐标系后为不同group类别间添加轴空白？

需求说明 #

你需要在使用coord_flip()翻转坐标系后，为group变量的三个类别组（种族、性别、地区）之间添加空白，当前已通过facet_wrap(~type)分面，并按组设置了填充色。

方法一：插入空因子水平（快速实现） #

通过在原有group的顺序中插入空字符串作为占位水平，即可在类别组之间留出空白，同时隐藏空水平的标签：

# 修改顺序，在每个类别组之间插入空水平
order <- c("white", "black", "hispanic", "", "female", "male", "", "north", "south")

ggplot(d, aes(x = group, y = percent, fill = group)) +
  geom_bar(stat = "identity") + 
  facet_wrap(~type) +
  coord_flip() +
  scale_x_discrete(
    limits = order,
    # 隐藏空水平的标签
    labels = function(x) ifelse(x == "", "", x)
  ) +
  scale_fill_manual(values = c(
    "white" = "skyblue", "black" = "skyblue", "hispanic" = "skyblue",
    "male" = "indianred1", "female" = "indianred1",
    "north" = "goldenrod2", "south" = "goldenrod2"
  )) +
  theme_minimal() +
  theme(
    legend.position = "none",
    # 调整轴文本间距，强化空白效果
    axis.text.y = element_text(margin = margin(r = 5))
  )

方法二：使用ggh4x嵌套坐标轴（专业分组显示） #

如果需要同时显示类别组的标题，推荐使用ggh4x包的嵌套坐标轴功能，自动在组间添加空白并展示分组标签：

1. 先安装并加载依赖包：

install.packages("ggh4x")
library(ggh4x)
library(dplyr)

2. 预处理数据，添加类别组变量并创建组合因子：

d <- d %>%
  mutate(
    # 为每个group分配所属类别组
    category = case_when(
      group %in% c("white", "black", "hispanic") ~ "种族",
      group %in% c("female", "male") ~ "性别",
      group %in% c("north", "south") ~ "地区"
    ),
    # 创建包含类别组和group的组合变量
    group_cat = interaction(category, group, sep = "||"),
    # 按原有顺序设置因子水平
    group_cat = factor(group_cat, levels = c(
      "种族||white", "种族||black", "种族||hispanic",
      "性别||female", "性别||male",
      "地区||north", "地区||south"
    ))
  )

3. 绘制嵌套轴图表：

ggplot(d, aes(x = group_cat, y = percent, fill = group)) +
  geom_bar(stat = "identity") + 
  facet_wrap(~type) +
  coord_flip() +
  scale_x_discrete(
    # 使用嵌套轴，按分隔符拆分组合变量
    guide = guide_axis_nested(delim = "||"),
    # 只显示group的标签
    labels = function(x) gsub(".*\\|\\|", "", x)
  ) +
  scale_fill_manual(values = c(
    "white" = "skyblue", "black" = "skyblue", "hispanic" = "skyblue",
    "male" = "indianred1", "female" = "indianred1",
    "north" = "goldenrod2", "south" = "goldenrod2"
  )) +
  theme_minimal() +
  theme(
    legend.position = "none",
    # 设置类别组标签的样式
    axis.text.y.top = element_text(face = "bold", size = 12)
  )

说明 #

• 方法一适合快速实现空白需求，无需额外包；
• 方法二能同时呈现分组结构，让图表逻辑更清晰，适合正式可视化场景。

内容的提问来源于stack exchange，提问作者a_todd12