在R中使用ggplot绘制专利数据嵌套环形图的技术求助

以下是针对你提出的两种嵌套环形图需求的完整实现方案，同时解决布局控制和标签添加问题，先修正原数据处理中的统计错误：

library(tidyverse)

# 加载并预处理数据
load(url("https://github.com/aterhorst/data/blob/master/patents.Rdata?raw=true"))

class_breakdown <- lens_reformatted %>%
  select(class = value, country_code = `Country Code`, country = `Country Name`) %>%
  mutate(
    country_code = ifelse(is.na(country), "XX", country_code),
    country = ifelse(country_code == "XX", "Unknown", country)
  ) %>%
  # 统计每个类别-国家组合的专利数，直接用n()统计分组内的条目数
  group_by(class, country) %>%
  summarise(patents = n(), .groups = "drop")

方案一：内层展示专利类别占比，外层按国家细分对应类别 #

实现逻辑 #

内层环展示每个专利类别的总数量占比，外层环将每个类别拆分为对应国家的专利数，通过计算每个分段的起始/中点位置实现标签定位：

# 1. 预处理内层（类别）数据，计算标签中点位置
inner_class_data <- class_breakdown %>%
  group_by(class) %>%
  summarise(total = sum(patents), .groups = "drop") %>%
  mutate(
    y_start = cumsum(lag(total, default = 0)),
    y_end = cumsum(total),
    y_mid = (y_start + y_end) / 2  # 类别标签的中心位置
  )

# 2. 预处理外层（国家）数据，计算每个国家在对应类别中的位置
outer_class_data <- class_breakdown %>%
  left_join(inner_class_data %>% select(class, y_start), by = "class") %>%
  group_by(class) %>%
  mutate(
    country_start = y_start + cumsum(lag(patents, default = 0)),
    country_end = y_start + cumsum(patents),
    country_mid = (country_start + country_end) / 2  # 国家标签的中心位置
  ) %>%
  ungroup()

# 3. 绘制嵌套环形图
ggplot() +
  # 内层类别环：x=1控制环的半径，width控制环的宽度
  geom_col(
    data = inner_class_data,
    aes(x = 1, y = total, fill = class),
    width = 0.8
  ) +
  # 外层国家环：x=2控制外层半径，白色边框区分不同国家
  geom_col(
    data = outer_class_data,
    aes(x = 2, y = patents, fill = class),
    width = 0.8,
    color = "white", size = 0.1
  ) +
  # 添加内层类别标签：x=0.6控制标签在内环内侧
  geom_text(
    data = inner_class_data,
    aes(x = 0.6, y = y_mid, label = class),
    size = 3, fontface = "bold"
  ) +
  # 添加外层国家标签：过滤专利数≥10的国家避免标签重叠，x=2.4控制标签在外环外侧
  geom_text(
    data = outer_class_data %>% filter(patents >= 10),
    aes(x = 2.4, y = country_mid, label = country),
    size = 2.5, hjust = 0
  ) +
  scale_fill_aaas() +
  coord_polar(theta = "y", start = 0) +
  # 控制x轴范围，避免标签超出画布
  scale_x_continuous(limits = c(0, 3)) +
  theme_void() +
  theme(legend.position = "none")

方案二：内层展示国家专利占比，外层细分专利类别 #

实现逻辑 #

内层环展示每个国家的总专利数占比，外层环将每个国家拆分为对应类别的专利数，同样通过计算分段中点定位标签：

# 1. 预处理内层（国家）数据，过滤少量专利的国家避免标签拥挤
inner_country_data <- class_breakdown %>%
  group_by(country) %>%
  summarise(total = sum(patents), .groups = "drop") %>%
  mutate(
    y_start = cumsum(lag(total, default = 0)),
    y_end = cumsum(total),
    y_mid = (y_start + y_end) / 2
  ) %>%
  filter(total >= 20)  # 仅保留专利数≥20的国家

# 2. 预处理外层（类别）数据，计算每个类别在对应国家中的位置
outer_country_data <- class_breakdown %>%
  inner_join(inner_country_data %>% select(country, y_start), by = "country") %>%
  group_by(country) %>%
  mutate(
    class_start = y_start + cumsum(lag(patents, default = 0)),
    class_end = y_start + cumsum(patents),
    class_mid = (class_start + class_end) / 2
  ) %>%
  ungroup()

# 3. 绘制嵌套环形图
ggplot() +
  # 内层国家环
  geom_col(
    data = inner_country_data,
    aes(x = 1, y = total, fill = country),
    width = 0.8
  ) +
  # 外层类别环
  geom_col(
    data = outer_country_data,
    aes(x = 2, y = patents, fill = class),
    width = 0.8,
    color = "white", size = 0.1
  ) +
  # 添加内层国家标签
  geom_text(
    data = inner_country_data,
    aes(x = 0.6, y = y_mid, label = country),
    size = 3, fontface = "bold"
  ) +
  # 添加外层类别标签：过滤专利数≥5的条目
  geom_text(
    data = outer_country_data %>% filter(patents >= 5),
    aes(x = 2.4, y = class_mid, label = class),
    size = 2.5, hjust = 0
  ) +
  scale_fill_viridis_d(option = "plasma") +
  coord_polar(theta = "y", start = 0) +
  scale_x_continuous(limits = c(0, 3)) +
  theme_void() +
  theme(legend.position = "bottom")

关键说明 #

• 环的半径由geom_col的x参数控制（如x=1、x=2），环的宽度由width参数调整
• 标签定位的核心是通过累加前序分段的数值，计算每个分段的中点位置，确保标签居中显示
• 由于国家数量多达119个，必须通过filter过滤少量专利的条目，否则标签会严重重叠，可根据需求调整阈值
• 原代码中sum(n())是错误用法：group_by后直接用n()即可统计每组的专利数量

内容的提问来源于stack exchange，提问作者aterhorst