如何在ggplot2柱状图中换行均值分离标签并居中于柱顶？

处理保密模拟数据时，用ggplot2绘制分组柱状图展示数据差异，遇到均值分离字母（hsd标签）过长导致相邻柱子标签重叠的问题。尝试用str_wrap函数换行但未成功，希望实现标签换行、居中对应柱顶且互不干扰。

用户提供的原始代码：

library(ggplot2)
library(tidyverse)

## Data to work with
data <- data.frame("trt" = c('t1',
                             't1',
                             't1',
                             't1',
                             't2',
                             't2',
                             't2',
                             't2',
                             't3',
                             't3',
                             't3',
                             't3',
                             't4',
                             't4',
                             't4',
                             't4'),
                   "splt" = c(1,
                              2,
                              3,
                              4,
                              1,
                              2,
                              3,
                              4,
                              1,
                              2,
                              3,
                              4,
                              1,
                              2,
                              3,
                              4),
                   "response"=c(82,
                                62,
                                17,
                                93,
                                55,
                                99,
                                42,
                                67,
                                56,
                                8,
                                62,
                                69,
                                38,
                                80,
                                86,
                                26),
                   standard_error=c(2,
                                    5,
                                    3,
                                    3,
                                    1,
                                    4,
                                    4,
                                    6,
                                    4,
                                    4,
                                    1,
                                    5,
                                    5,
                                    5,
                                    6,
                                    4),
                   hsd=c('abcdef',
                         'abcdef',
                         'abcdefg',
                         'abcdefg',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh',
                         'abcdefgh'))

# Change to factor
data$splt <- as.factor(data$splt)

ggplot(data, aes(fill = splt, x = trt, y = response))+
  geom_bar(stat="identity", width = 0.6, position = position_dodge(width = 0.8), col = "black")+
  geom_errorbar(aes(ymin = response, ymax = response + standard_error), 
                width = 0.3, position = position_dodge(0.8)) +
  guides(fill = guide_legend("Splt"))+
  xlab("Treatment")+
  ylab('Response')+
  ggtitle("Title",
          subtitle = "Subtitle")+
  theme(plot.title = element_text(hjust=0.5, size = 15),
        plot.subtitle = element_text(hjust=0.5, size = 11),
        axis.text = element_text(size = 10),
        axis.title = element_text(size = 12),
        legend.key.size = unit(0.3, 'cm'))+
  geom_text(aes(label = str_wrap(hsd, width = 10), y = response + standard_error, group = splt), vjust = -0.9, 
            size = 4, position = position_dodge(width = 0.8))+
  ylim(0, 110)+
  scale_x_discrete(labels = function(x) str_wrap(x, width = 10))

问题核心是str_wrap的宽度设置过大，且标签的垂直位置、对齐方式未适配换行需求。以下是修改后的代码：

library(ggplot2)
library(tidyverse)

# 整理后的数据集
data <- data.frame(
  trt = rep(c('t1', 't2', 't3', 't4'), each = 4),
  splt = as.factor(rep(1:4, 4)),
  response = c(82,62,17,93,55,99,42,67,56,8,62,69,38,80,86,26),
  standard_error = c(2,5,3,3,1,4,4,6,4,4,1,5,5,5,6,4),
  hsd = c('abcdef','abcdef','abcdefg','abcdefg',
          rep('abcdefgh', 12))
)

ggplot(data, aes(fill = splt, x = trt, y = response))+
  geom_bar(stat="identity", width = 0.6, position = position_dodge(width = 0.8), col = "black")+
  geom_errorbar(aes(ymin = response, ymax = response + standard_error), 
                width = 0.3, position = position_dodge(0.8)) +
  guides(fill = guide_legend("Splt"))+
  xlab("Treatment")+
  ylab('Response')+
  ggtitle("Title", subtitle = "Subtitle")+
  theme(
    plot.title = element_text(hjust=0.5, size = 15),
    plot.subtitle = element_text(hjust=0.5, size = 11),
    axis.text = element_text(size = 10),
    axis.title = element_text(size = 12),
    legend.key.size = unit(0.3, 'cm')
  )+
  # 调整标签的换行、位置与对齐
  geom_text(
    aes(
      label = str_wrap(hsd, width = 3),  # 缩小宽度强制换行
      y = response + standard_error + 2, # 增加垂直空间避免与误差线重叠
      group = splt
    ),
    vjust = 0,          # 标签从误差线顶部向下排版
    size = 3.5,         # 缩小字号减少占用空间
    position = position_dodge(width = 0.8),
    lineheight = 0.8    # 调整行间距让换行标签更紧凑
  )+
  ylim(0, 120)+         # 扩大y轴上限，避免标签被截断
  scale_x_discrete(labels = function(x) str_wrap(x, width = 10))

关键修改点 #

• str_wrap(width = 3)：根据标签长度设置更小的宽度，强制长文本按指定字符数换行
• 调整标签y位置：通过response + standard_error + 2给标签预留足够垂直空间，避免与误差线重叠
• vjust = 0：改变标签垂直对齐方向，让标签从误差线顶部向下排列，避免高柱子的标签与上方内容重叠
• lineheight = 0.8：缩小行间距，减少换行标签的垂直占用
• 扩大y轴上限：将ylim(0,110)改为ylim(0,120)，确保换行后的标签不会被图框截断

如果仍有重叠，可进一步尝试：

• 继续缩小标签字号（如size = 3）
• 增大position_dodge的width值，拉开分组柱子的间距
• 用ifelse动态调整不同高度柱子的标签y偏移量

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