如何在ggplot2填充条形图内显示vacc_status_broad类别而非仅图例

要把vacc_status_broad的类别标签直接显示在填充条形内部，核心是用geom_text()配合position_fill()来定位文本，同时可按需隐藏原图例。

示例修改代码 #

假设你的原始代码和数据如下（基于常见场景模拟，可直接替换为你的实际数据）：

library(ggplot2)

# 模拟你提供的dput数据结构
df <- structure(list(
  species = rep("Dog", 9),
  group = rep(c("Group A", "Group B", "Group C"), each = 3),
  vacc_status_broad = rep(c("Unvaccinated", "Partially Vaccinated", "Fully Vaccinated"), 3),
  count = c(12, 24, 36, 18, 30, 42, 8, 16, 48)
), class = "data.frame", row.names = c(NA, -9L))

# 原始绘图代码（仅图例显示类别）
ggplot(df, aes(x = group, y = count, fill = vacc_status_broad)) +
  geom_bar(stat = "identity", position = "fill") +
  labs(title = "Dog Vaccination Status", x = "Group", y = "Proportion")

修改后实现内部标签的代码：

ggplot(df, aes(x = group, y = count, fill = vacc_status_broad)) +
  geom_bar(stat = "identity", position = "fill") +
  # 添加居中的类别标签，position_fill与条形填充逻辑匹配
  geom_text(aes(label = vacc_status_broad),
            position = position_fill(vjust = 0.5),
            color = "white", # 文本颜色适配深色填充，可按需调整
            size = 3) +
  labs(title = "Dog Vaccination Status", x = "Group", y = "Proportion") +
  theme(legend.position = "none") # 隐藏原图例

关键细节说明 #

• position_fill(vjust=0.5)：确保文本在每个填充分段的垂直居中位置显示，和条形的填充比例完全对齐。
• 文本样式可灵活调整：比如把color改成"black"适配浅色填充，或调整size改变字号。
• 如果需要同时显示比例数值，可把label改成paste0(vacc_status_broad, "\n", scales::percent(count/sum(count)))（需先按分组计算比例，或用stat="count"配合aes(label=..prop..)）。

进阶优化：自动适配文本颜色 #

如果填充色有深浅差异，可自定义函数根据背景亮度切换文本颜色：

# 自定义函数：根据背景色亮度选择黑/白文本
text_color <- function(fill_col) {
  rgb_vals <- col2rgb(fill_col)
  luminance <- (0.299*rgb_vals[1] + 0.587*rgb_vals[2] + 0.114*rgb_vals[3])/255
  ifelse(luminance > 0.5, "black", "white")
}

# 生成适配的文本颜色列
df$text_col <- text_color(df$vacc_status_broad)

ggplot(df, aes(x = group, y = count, fill = vacc_status_broad)) +
  geom_bar(stat = "identity", position = "fill") +
  geom_text(aes(label = vacc_status_broad, color = I(text_col)),
            position = position_fill(vjust = 0.5),
            size = 3) +
  labs(title = "Dog Vaccination Status", x = "Group", y = "Proportion") +
  theme(legend.position = "none")

内容的提问来源于stack exchange，提问作者Rakesh Chand