如何在ggplot2中按样本量缩放多个饼图的大小？

问题说明 #

现有R代码通过ggplot2绘制了各学科男女占比的分面饼图，需要将饼图按照对应学科的总学生数（数学158人最大，英语31人最小）从大到小调整尺寸，同时保持分面顺序与总人数排序一致。

修正后的完整代码 #

library(dplyr)
library(ggplot2)

# 数据处理：计算总人数并排序学科
df <- data.frame(Subjects = c("Literature", "Literature", "Geography", "Geography", "Arts", "Arts", 
                              "English", "English", "Math", "Math", "History", "History", 
                              "Biology", "Biology", "French", "French"),
                 Sex = rep(c("Girls", "Boys"), 8), # 修正原代码重复次数错误
                 N = c(15, 20, 30, 22, 77, 40, 14, 17, 
                       80, 78, 52, 65, 32, 56, 89, 41)) %>%
  group_by(Subjects) %>%
  mutate(total_N = sum(N), # 计算每个学科总人数
         pct = proportions(N) * 100) %>%
  mutate(pct = round(pct, 1)) %>%
  ungroup() %>%
  # 将学科按总人数降序转为有序因子，控制分面顺序
  mutate(Subjects = factor(Subjects, 
                           levels = arrange(., desc(total_N)) %>% pull(Subjects) %>% unique()))

# 绘制带尺寸差异的分面饼图
df %>%
  ggplot(aes(x = total_N, y = pct, fill = Sex)) +
  geom_bar(stat = "identity", width = 1, color = "white", na.rm = TRUE) +
  coord_polar("y", start = 0) +
  facet_wrap(~ Subjects,
             ncol = 4,
             strip.position = "bottom",
             scales = "free_x") + # 允许x轴自由缩放，实现饼图尺寸差异
  geom_text(aes(y = pct, label = paste0(pct, "%")), size = 5, 
            position = position_stack(vjust = 0.5)) +
  theme_void() +
  theme(legend.title = element_blank(),
        strip.text = element_text(size = 12),
        plot.title = element_text(hjust = 0.0),
        legend.text = element_text(size = 12),
        legend.spacing.y = unit(0.0, 'cm'),
        legend.box.spacing = unit(-2, "pt")) + 
  guides(fill = guide_legend(byrow = TRUE)) + 
  scale_fill_brewer(palette = "Set2") +
  scale_x_continuous(limits = c(0, NA)) # 确保x轴从0起始，避免饼图偏移

关键修改点 #

• 修正数据错误：原代码中rep(c("Girls", "Boys"), 2)无法匹配16行数据，改为rep(c("Girls", "Boys"), 8)；
• 新增总人数字段：通过total_N = sum(N)计算每个学科的总学生数；
• 控制分面顺序：将Subjects转为按total_N降序排列的有序因子，让分面从总人数最多的学科开始；
• 实现尺寸差异：将x轴映射为total_N，结合facet_wrap的scales = "free_x"参数，让每个分面的x轴范围独立，极坐标下x轴长度对应饼图半径，从而实现饼图大小与总人数成正比；
• 优化饼图位置：添加scale_x_continuous(limits = c(0, NA))确保x轴从0开始，避免饼图出现偏移。

内容的提问来源于stack exchange，提问作者always.learning