ggplot柱状图：实现分组绝对值与子组占比显示的问题

需求 #

• 每个柱状顶部显示对应分组的绝对值总计数
• 每个柱状内按子组展示占比分布，并在子组区域内显示占比数值

现有尝试与问题 #

1. 数据处理代码 #

df <- 
df %>% 
 group_by(directed_to_whom, treatment_details) %>%   
  mutate(direct_period_count=n()) %>% # 按treatment_details分组下各directed_to_whom的计数
  group_by(treatment_details) %>% 
  mutate(period_count=n())  %>% # 每个treatment_details分组的总计数
  mutate(directed_to_whom_percentage = direct_period_count/period_count*100) # 计算占比

该代码无报错，但生成的数据存在重复行，导致绘图逻辑混乱。

2. 初始绘图代码 #

ggplot(experimentgraph_directed_to_whom, aes(x = treatment_details, y = period_count, 
                       fill = directed_to_whom, label=directed_to_whom_percentage))+
      geom_bar(position = "fill", stat = "identity") +
  scale_y_continuous(labels = scales::label_percent(accuracy = 1) + 
  geom_text(position = position_stack(vjust = 0.5), size = 4)) +
  theme(plot.title = element_text(size = 18, face = "bold")) +
  theme_bw()+
    theme(plot.title = element_text(size = 5, face = "bold"),
          axis.text.x = element_text(angle =90, vjust = 0.5)) 

存在语法错误（scale_y_continuous内嵌套geom_text），且未正确绑定占比数据，导致无法显示子组占比和顶部绝对值。

3. 修改后的尝试 #

数据处理：

df <- 
  experimentgraph_directed_to_whom %>%
  distinct()

绘图代码：

ggplot(df, aes(x = treatment_details, y = period_count, 
                                             fill = directed_to_whom, label=directed_to_whom_percentage))+
  geom_bar(position = "fill", stat = "identity") +
  geom_text(aes(y = 1,label = sum(direct_period_count)), vjust = 0) +  
  geom_text(aes(label = scales::label_percent(0.1, scale = 1) (directed_to_whom_percentage)), position = position_fill(.5)) + 
  scale_y_continuous(labels = scales::label_percent(accuracy = 1) + 
                       geom_text(position = position_stack(vjust = 0.5), size = 4)) +
  theme(plot.title = element_text(size = 18, face = "bold")) +
  theme_bw()+
  theme(plot.title = element_text(size = 5, face = "bold"),
        axis.text.x = element_text(angle =90, vjust = 0.5)) 

仍未解决数据重复、语法错误及文本定位问题，无法实现预期效果。

正确解决方案 #

步骤1：准确聚合数据 #

先通过聚合得到每个分组-子组的计数及总计数，避免重复行：

library(dplyr)
library(ggplot2)
library(scales)

df_summary <- experimentgraph_directed_to_whom %>%
  # 统计每个treatment_details下各directed_to_whom的数量
  count(treatment_details, directed_to_whom, name = "direct_period_count") %>%
  # 按treatment_details分组，计算总计数和占比（小数形式）
  group_by(treatment_details) %>%
  mutate(period_count = sum(direct_period_count),
         percentage = direct_period_count / period_count)

步骤2：绘制目标柱状图 #

ggplot(df_summary, aes(x = treatment_details, y = percentage, fill = directed_to_whom)) +
  # 堆叠占比柱状图
  geom_bar(stat = "identity", position = "fill") +
  # 子组内显示占比文本，居中对齐
  geom_text(aes(label = percent(percentage, accuracy = 1)),
            position = position_fill(vjust = 0.5), size = 4) +
  # 柱子顶部显示总计数绝对值，偏移至柱子上方
  geom_text(data = df_summary %>% distinct(treatment_details, period_count),
            aes(x = treatment_details, y = 1.02, label = period_count),
            vjust = 0, size = 4.5, fontface = "bold") +
  # y轴设置为百分比格式
  scale_y_continuous(labels = percent_format(accuracy = 1)) +
  # 主题调整
  theme_bw() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1),
        plot.title = element_text(size = 14, face = "bold")) +
  # 自定义标签
  labs(title = "分组占比与总计数分布", x = "分组", y = "占比", fill = "子组")

关键修正说明 #

• 数据层面：用count()直接聚合，避免原代码中重复行导致的绘图错误；占比用小数计算，适配ggplot的百分比格式化。
• 绘图逻辑：
  • y轴绑定percentage，配合position="fill"实现占比堆叠，确保所有柱子高度统一为100%。
  • 顶部绝对值文本单独提取每个分组的唯一总计数，设置y=1.02避免与柱子重叠。
  • 修正原代码中scale_y_continuous嵌套图层的语法错误，分离各绘图元素。

内容的提问来源于stack exchange，提问作者nesta1990