如何为geom_segment应用position_dodge实现分组箭头无重叠？

问题描述 #

需要绘制一款以箭头形式呈现的修改版条形图，手动指定每个箭头的起止值。当前使用geom_segment绘制的箭头存在重叠，要求按国家分组（每个国家对应数据框category列的5个箭头），实现同组箭头并排展示，且国家间保留较宽间距。

数据集 #

consumption_supplychain <- structure(list(country = c("Europe", "Europe", "Europe", "Europe", "Europe", "China", "China", "China", "China", "China", "India", "India", "India", "India", "India", "Brazil", "Brazil", "Brazil", "Brazil", "Brazil", "USA", "USA", "USA", "USA", "USA"), category = c(1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L, 1L, 2L, 3L, 4L, 5L), y = c(0L, 899620L, 783001L, 1073611L, 875287L, 0L, 2397951L, 3529763L, 4178099L, 4352804L, 0L, 1290543L, 1316816L, 1377391L, 1272729L, 0L, 963706L, 961357L, 544286L, 276219L, 0L, 1051162L, 441454L, 262085L, 348517L), yend = c(899620L, 783001L, 1073611L, 875287L, 1442741L, 2397951L, 3529763L, 4178099L, 4352804L, 2564677L, 1290543L, 1316816L, 1377391L, 1272729L, 923000L, 963706L, 961357L, 544286L, 276219L, 249263L, 1051162L, 441454L, 262085L, 348517L, 1314709L)), row.names = c(NA, -25L), class = c("data.table", "data.frame"))

过往尝试的代码（存在重叠/混乱问题） #

# 尝试用position_dodge但箭头仍重叠
ggplot(
    aes(
      x = country, xend = country, 
      y = y, yend = yend,
      ),
    position = position_dodge(width = 0.2)) +
  geom_segment(
    arrow = arrow(angle = 30, length = unit(0.2, 'cm')),
    linewidth = 1
  )

# 尝试用geom_arrowsegment加jitter导致位置混乱
ggplot(consumption_supplychain) +
  geom_arrowsegment(aes(x = country, xend = country, y = y, yend = yend),
                    position = position_jitter())

解决方案 #

核心思路是给每个国家下的不同category分配带水平偏移的x坐标，既实现同组箭头并排，又保留国家间的宽间距。

步骤1：预处理数据，生成偏移x值 #

library(ggplot2)
library(data.table)

# 将country转为因子，固定顺序
consumption_supplychain[, country_factor := factor(country, levels = unique(country))]
# 计算每个category的x偏移：国家基础位置 + 小偏移，让5个category在国家中心左右均匀分布
consumption_supplychain[, x_offset := as.numeric(country_factor) + (category - 3)/10]

步骤2：绘制箭头图 #

ggplot(consumption_supplychain, aes(x = x_offset, xend = x_offset, y = y, yend = yend)) +
  geom_segment(arrow = arrow(angle = 30, length = unit(0.2, 'cm')), linewidth = 1) +
  # 重置x轴刻度为国家名称，对应每个国家的中心位置
  scale_x_continuous(
    breaks = unique(as.numeric(consumption_supplychain$country_factor)),
    labels = unique(consumption_supplychain$country)
  ) +
  labs(x = "国家", y = "数值") +
  theme_bw()

说明 #

• (category - 3)/10的偏移量可根据需求调整：数值越小，同组箭头间距越窄；反之则越宽
• 若不用data.table，也可改用dplyr处理数据，逻辑一致
• 可添加color = factor(category)映射颜色，更清晰区分不同category的箭头

内容的提问来源于stack exchange，提问作者msug