如何调整ggplot饼图标签位置：解决停靠型自行车标签可读性问题

解决方案 #

核心原因

position_stack在极坐标下的行为和直角坐标不同，直接调整vjust无法实现垂直方向的位置偏移，需要通过计算饼块中心位置或调整x轴参数来实现需求。

1. 单独调整docked bike标签位置

先在数据处理阶段计算每个饼块的中心位置，再针对docked bike标签单独设置y轴偏移量：

数据处理（补充位置计算）

library(tidyverse)
library(scales)

user_type_ride_type <- clean_rides %>% 
  group_by(rideable_type, member_casual) %>% 
  summarise(number_of_rides = n(),
            total_ride_time = sum(ride_length),
            average_ride_time = mean(ride_length)) %>%
  mutate(percentage = case_when(
    member_casual == "casual" ~ round(number_of_rides / sum(number_of_rides) * 100, 0),
    member_casual == "member" ~ round(number_of_rides / sum(number_of_rides) * 100, 0)
  )) %>%
  ungroup()

casual_data <- user_type_ride_type %>% 
  filter(member_casual == "casual") %>%
  arrange(rideable_type) %>%
  # 计算累积骑行次数，推导每个饼块的中心位置
  mutate(cum_rides = cumsum(number_of_rides),
         center = cum_rides - number_of_rides/2)

绘图代码（调整单个标签）

ggplot(casual_data, aes(x = "", y = number_of_rides, fill = rideable_type)) +
  geom_bar(stat = "identity", width = 1) +
  scale_fill_manual(values = c("#00C19F", "#D7EF16", "#9100B1"),
                    labels = c("classic bike", "electric bike", "docked bike")) +
  coord_polar("y", start=0) +
  # 绘制普通标签
  geom_text(data = filter(casual_data, rideable_type != "docked bike"),
            aes(label = comma(number_of_rides), y = center),
            size = 4) +
  # 单独调整docked bike标签，通过y轴偏移上移（偏移量根据数据量级调整）
  geom_text(data = filter(casual_data, rideable_type == "docked bike"),
            aes(label = comma(number_of_rides), y = center + 50000),
            size = 4) +
  labs(title = "Total Rides by Bike Type for Casual Users",
       fill = "Bike Type",
       caption = "Data Source: Motivate International Inc") +
  theme_void() +
  theme(legend.position = "bottom",
        plot.caption.position = "plot",
        plot.caption = element_text(color = "gray", hjust = 3.8, vjust = 1))

2. 将所有标签移至饼图外缘

通过设置标签的x值大于1（饼图默认x轴范围是0-1），配合position_stack对齐饼块中心线，即可将标签移到外侧：

绘图代码

ggplot(casual_data, aes(x = "", y = number_of_rides, fill = rideable_type)) +
  geom_bar(stat = "identity", width = 1) +
  scale_fill_manual(values = c("#00C19F", "#D7EF16", "#9100B1"),
                    labels = c("classic bike", "electric bike", "docked bike")) +
  coord_polar("y", start=0) +
  # x=1.2将标签放在饼图外侧，vjust=0.5确保标签对齐饼块中心
  geom_text(aes(label = comma(number_of_rides), x = 1.2),
            position = position_stack(vjust = 0.5),
            size = 4) +
  labs(title = "Total Rides by Bike Type for Casual Users",
       fill = "Bike Type",
       caption = "Data Source: Motivate International Inc") +
  theme_void() +
  theme(legend.position = "bottom",
        plot.caption.position = "plot",
        plot.caption = element_text(color = "gray", hjust = 3.8, vjust = 1))

关键说明 #

• 调整单个标签时，center + 50000中的偏移量需要根据你的数据量级调整，数值越大标签上移越多
• 移至外缘时，x=1.2可根据饼图大小修改，值越大标签离饼图越远；position_stack(vjust=0.5)保证标签在每个饼块的中心线上

内容的提问来源于stack exchange，提问作者Marinda_L