如何用ggplot优雅绘制因子水平均值的条形图顶部线段？

问题描述 #

我需要绘制因子各水平的均值，但不想使用geom_col绘制完整条形图，仅展示条形图的顶部线段。以下代码可实现需求，但不够优雅（尤其是硬编码breaks参数，以及在scale_x_continuous调用中引用the_means$group来设置标签的部分）。请问是否有简洁方法让ggplot为每个因子水平绘制线段？

原实现代码 #

blah <- 
  data.frame(
    group = factor(c(rep("A", 3), rep("B", 5))), 
    value = c(1, 2, 3, 7, 8, 9, 10, 12)
  )

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
the_means <- 
  blah |> 
  summarise(
    group_mean = mean(value),
    .by = group
  ) |> 
  mutate(
    x_start = as.numeric(group) - .3,
    x_end = as.numeric(group) + .3,
  )

library(ggplot2)
ggplot() +
  geom_segment(
    aes(x = x_start, xend = x_end, y = group_mean, yend = group_mean),
    data = the_means
  ) +
  scale_x_continuous(
    breaks = c(1, 2), minor_breaks = NULL, labels = the_means$group
  ) +
  expand_limits(y = 0)

简洁优化方案 #

可以利用ggplot的离散x轴特性，结合geom_errorbar或stat_summary实现，无需手动处理刻度和线段位置：

方法1：提前计算均值，用geom_errorbar绘制线段 #

library(ggplot2)
library(dplyr)

blah <- 
  data.frame(
    group = factor(c(rep("A", 3), rep("B", 5))), 
    value = c(1, 2, 3, 7, 8, 9, 10, 12)
  )

# 计算各分组均值
the_means <- blah |> 
  summarise(group_mean = mean(value), .by = group)

# 绘制顶部线段
ggplot(the_means, aes(x = group, y = group_mean)) +
  geom_errorbar(aes(ymin = group_mean, ymax = group_mean), width = 0.6) + # width控制线段长度
  expand_limits(y = 0)

方法2：无需提前计算，用stat_summary一步到位 #

library(ggplot2)

blah <- 
  data.frame(
    group = factor(c(rep("A", 3), rep("B", 5))), 
    value = c(1, 2, 3, 7, 8, 9, 10, 12)
  )

ggplot(blah, aes(x = group, y = value)) +
  stat_summary(
    fun = mean, # 指定计算均值
    geom = "errorbar", # 使用errorbar几何对象
    aes(ymin = after_stat(y), ymax = after_stat(y)), # 让上下限等于均值，只显示水平线段
    width = 0.6 # 控制线段长度
  ) +
  expand_limits(y = 0)

优化优势 #

• 直接使用离散x轴，ggplot自动处理刻度标签，无需硬编码breaks和手动关联标签
• 无需手动转换因子为数值计算线段起止位置，代码更简洁易读
• 方法2省去了提前汇总数据的步骤，逻辑更连贯

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