R语言ggplot2饼图标签外移、图例调整及可视化优化咨询

1. 扇区数值标签外移 #

极坐标下x值越大越靠近画布外缘，只需修改geom_text的x映射值为大于1的数值即可将标签外移，额外添加check_overlap = TRUE参数可自动隐藏重叠的标签，修改示例：

geom_text(aes(x = 1.3, label = value), size=4, 
          position = position_stack(vjust = 0.5), alpha = 1, color = "black",
          check_overlap = TRUE)

如果需要完全避免标签重叠，可使用ggrepel包的geom_text_repel，会自动为重叠标签添加引导线。

2. 图例位置上移 #

通过theme中的legend.position参数配置图例相对位置，坐标取值范围为0-1，(0,0)对应画布左下角，(1,1)对应画布右上角。如需将图例移动到右上角区域，添加如下配置即可：

theme(
  legend.position = c(0.9, 0.95), 
  legend.justification = "top",
  legend.box.just = "right"
)

可根据实际显示效果微调坐标数值。

3. 可视化优化建议 #

• 改用环形饼图（甜甜圈图）：将geom_col(width = 1)修改为width = 0.6，中间空白区域可添加总内存占用数值，视觉上更清爽，也能留出更多空间放标签
• 处理过长图例文本：用stringr::str_wrap()对Legend字段自动换行，避免图例横向占用过多空间，比如fill = fct_inorder(str_wrap(Legend, 25))
• 移除多余视觉元素：在theme中添加panel.grid = element_blank()去掉极坐标网格线，降低视觉干扰
• 优化标签可读性：数值标签添加单位，改为label = paste0(value, "MB")
• 如重点是对比各类别数值大小，建议改用水平条形图，对小占比类别的数值差异展示会比饼图清晰很多

df <- structure(list(group = c("Other", "Pool", "VNC", "File", "EtwB", "Thre", "HTab", "Vkmc", "Ntfx", "MmCa"), value = c(20.32, 6.16, 4.03, 2.2, 2.05, 2.01, 1.72, 1.7, 1.56, 1.27), FileName = c("", NA, "C:\\Windows\\System32\\Drivers\\netvsc.sys", NA, NA, NA, NA, "C:\\Windows\\System32\\Drivers\\vmbkmcl.sys", NA, NA), Description = c("", "<unknown> (Pool tables, etc.)", "Virtual NDIS Miniport", "<unknown> (File objects)", "nt!etw (Etw Buffer)", "nt!ps (Thread objects)", "<unknown> (Hash Table pool)", "Hyper-V VMBus KMCL", NA, "nt!mm (Mm control areas for mapped files)"), Legend = c("Other", "Pool - <unknown> (Pool tables, etc.)", "VNC - Virtual NDIS Miniport - C:\\Windows\\System32\\Drivers\\netvsc.sys", "File - <unknown> (File objects)", "EtwB - nt!etw (Etw Buffer)", "Thre - nt!ps (Thread objects)", "HTab - <unknown> (Hash Table pool)", "Vkmc - Hyper-V VMBus KMCL - C:\\Windows\\System32\\Drivers\\vmbkmcl.sys", "Ntfx", "MmCa - nt!mm (Mm control areas for mapped files)")), row.names = c(NA, -10L), class = c("tbl_df", "tbl", "data.frame"))
df2 <- structure(list(group = c("Other", "Pool", "VNC", "File", "EtwB", "Thre", "HTab", "Vkmc", "Ntfx", "MmCa"), value = c(20.32, 6.16, 4.03, 2.2, 2.05, 2.01, 1.72, 1.7, 1.56, 1.27), Legend = c("Other", "Pool - <unknown> (Pool tables, etc.)", "VNC - Virtual NDIS Miniport - C:\\Windows\\System32\\Drivers\\netvsc.sys", "File - <unknown> (File objects)", "EtwB - nt!etw (Etw Buffer)", "Thre - nt!ps (Thread objects)", "HTab - <unknown> (Hash Table pool)", "Vkmc - Hyper-V VMBus KMCL - C:\\Windows\\System32\\Drivers\\vmbkmcl.sys", "Ntfx", "MmCa - nt!mm (Mm control areas for mapped files)"), csum = c(43.02, 22.7, 16.54, 12.51, 10.31, 8.26, 6.25, 4.53, 2.83, 1.27), pos = c(32.86, 19.62, 14.525, 11.41, 9.285, 7.255, 5.39, 3.68, 2.05, 0.635)), row.names = c(NA, -10L), class = c("tbl_df", "tbl", "data.frame"))

library(ggplot2)
library(tidyverse)
library(RColorBrewer)
# 如需使用带引线的标签可加载ggrepel
# library(ggrepel)

nb.cols <- 10
mycolors <- colorRampPalette(brewer.pal(8, "Dark2"), interpolate='spline')(nb.cols)
font_size <- 20

# 自动换行图例文本
df$Legend <- str_wrap(df$Legend, width = 25)

plot_to_save <- 
  ggplot(df, aes(x = "", y = value, fill = fct_inorder(Legend))) +
  # 改为环形饼图可把width改成0.6
  geom_col(width = 1, color = 1, alpha= 0.8) +
  # 标签外移，添加防重叠参数
  geom_text(aes(x = 1.3, label = paste0(value, "MB")), size=4, 
            position = position_stack(vjust = 0.5), alpha = 1, color = "black",
            check_overlap = TRUE) +
  # 如需用带引线的标签，替换上面的geom_text为下面的代码
  # geom_text_repel(aes(label = paste0(value, "MB")), size=4,
  #           position = position_stack(vjust = 0.5), segment.size = 0.3) +
  coord_polar(theta = "y") +
  guides(fill = guide_legend(title = "Driver")) +
  scale_y_continuous(breaks = df2$pos, labels = df$group) +
  theme_minimal() +
  theme(
    axis.ticks = element_blank(),
    axis.title = element_blank(),
    axis.text = element_text(size = font_size-4),
    panel.background = element_rect(fill = "white"),
    # 移除网格线
    panel.grid = element_blank(),
    text = element_text(family = "Segoe UI", size = font_size),
    # 调整图例到右上角
    legend.position = c(0.9, 0.95),
    legend.justification = "top",
    legend.box.just = "right"
  ) +
  scale_fill_manual(values = mycolors) +
  ggtitle("Memory NonPagedUsed, MB")

ggsave(
  file = "c:\\temp\\pie_chart.png",
  plot = plot_to_save,
  width = 16,
  height = 8,
  dpi = 150,
  units = "in",
  type = "cairo-png"
)

内容的提问来源于stack exchange，提问作者Maxim