如何用ggplot2绘制仅展示单模态的分组百分比条形图（R语言）

核心思路 #

要只展示某一关系状态（如"non stable"）在各年龄组的百分比，需先筛选出目标类别，再计算该类别在对应年龄组中的占比，最后绘制单一条形图而非堆叠图。

方法一：直接筛选目标类别后计算占比 #

library(tidyverse)
library(scales)

# 原始数据
age10 <- c("18-34","35-54", "55+", "55+","35-54","18-34","18-34","35-54","35-54","35-54")
relation <- c("stable","non stable", "non stable", "stable", "stable", "stable", "non stable", "non stable", "stable", "stable")

df <- data.frame(age10,relation)

# 数据处理：计算各年龄组中非稳定关系的百分比
df_single <- df %>%
  group_by(age10) %>%
  mutate(group_total = n()) %>%  # 统计每个年龄组的总人数
  filter(relation == "non stable") %>%  # 只保留非稳定关系的数据
  summarise(non_stable_percent = n()/group_total)  # 计算占比

# 绘图
ggplot(df_single) +
  aes(x = age10, y = non_stable_percent) +
  geom_bar(stat = "identity", fill = "#2ca02c") +  # 单一颜色填充条形
  scale_y_continuous(labels = percent) +  # 将数值转为百分比格式
  labs(x = "年龄组", y = "百分比", 
       title = "各年龄组非稳定关系占比")

方法二：基于原汇总数据调整 #

如果想保留原汇总逻辑，也可以在已分组统计的基础上筛选目标类别并计算占比：

# 基于原代码的汇总数据调整
df1 <- df %>%
  group_by(age10, relation) %>%
  summarise(n = n(), .groups = "drop") %>%
  group_by(age10) %>%
  mutate(total = sum(n), percent = n/total) %>%  # 计算每组总数及目标类别占比
  filter(relation == "non stable")  # 筛选目标类别

ggplot(df1) +
  aes(x = age10, y = percent) +
  geom_bar(stat = "identity", fill = "#d62728") +
  scale_y_continuous(labels = percent) +
  labs(x = "年龄组", y = "百分比", 
       title = "各年龄组非稳定关系占比")

关键修改点 #

• 筛选目标类别：通过filter(relation == "non stable")剔除不需要的关系状态数据
• 计算组内占比：按年龄组分组后统计总人数，再计算目标类别的占比，而非用position = "fill"生成堆叠图
• 简化绘图映射：移除fill = relation映射（仅展示单一类别），直接用固定颜色填充条形

内容的提问来源于stack exchange，提问作者lpambout