基于ggplot2的带个体与整体均值的箱线图绘制及排序问题

我需要绘制一幅箱线图，要求：

• 每个行为类别对应一个箱线，隐藏箱线的异常值
• 用红色大点显示该行为类别的整体均值
• 抖动显示数据集中9个个体在该行为下的均值（替换原来的原始数据抖动，避免图表混乱）

但修改代码后，出现两个问题：

• 行为类别无法按指定顺序排列
• 自定义填充颜色失效

初始代码 #

library(ggplot2)

ggplot(Seen2, aes(x=Behaviour, y=Roll_Avg, fill=Behaviour)) +
  geom_boxplot(outlier.shape= NA) +
  geom_point(aes(fill=Behaviour), size=2, position=position_jitter(width=0.2, height=0.1)) +
  stat_summary(fun=mean, geom="point", shape=20, size=5, color="red", fill="red") +
  theme_classic() + my_scale + 
  theme(axis.text.y=element_text(size=16, angle=0))+
  ylim(-30, 30) 

更新后（排序失效）的代码 #

my_colors <- c("#CCFFFF", "#000000", "#7F7F7F", "#336699", "#008080", "#00CCFF", "#264AE2")
names(my_colors) <- levels(factor(c((Seen2$Behaviour), levels(Seen2$Behaviour))))
my_scale <- scale_fill_manual(name="Behaviour", values=my_colors,)
behavssec$Behaviour <- factor(Seen2$Behaviour, 
                              levels=c("Burst", "High energy swimming", 
                                         "Medium energy swimming", 
                                         "Low energy swimming", "Travel",
                                         "Ascending", "Descending"))

ggplot(Seen2, aes(x=Behaviour, y=Roll_Avg, fill=Behaviour)) +
  geom_boxplot(outlier.shape= NA) +
  geom_point(data=means, size=2, position=position_jitter(width=0.2, height=0.1)) + 
  stat_summary(fun=mean, geom="point", shape=20, size=5, color="red", fill="red") +
  theme_classic() + my_scale + 
  theme( axis.text.y= element_text( size=16, angle =0)) +
  ylim(-30, 30) 

数据集 #

Seen2 <- structure(list(SharkID = c(9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 
9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 
9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 9L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 
5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 6L, 
6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 
6L, 6L, 6L, 6L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L), Behaviour = c("Low.energy.swimming", 
"Low.energy.swimming", "Low.energy.swimming", "Low.energy.swimming", 
"Low.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "Travel", "Travel", "Travel", "Travel", 
"Travel", "Burst", "Burst", "Burst", "Burst", "Burst", "Ascending", 
"Ascending", "Ascending", "Ascending", "Ascending", "Ascending", 
"Ascending", "Descending", "Descending", "Descending", "Descending", 
"Descending", "Descending", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "Medium.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "Medium.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "Medium.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "Medium.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "Burst", "Burst", "Burst", "Burst", "Burst", 
"Burst", "High.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "Medium.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "Medium.energy.swimming", "Medium.energy.swimming", 
"Low.energy.swimming", "Low.energy.swimming", "Low.energy.swimming", 
"Low.energy.swimming", "Low.energy.swimming", "Low.energy.swimming", 
"Low.energy.swimming", "Low.energy.swimming", "Low.energy.swimming", 
"Low.energy.swimming", "Low.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "High.energy.swimming", "High.energy.swimming", 
"High.energy.swimming", "Burst", "Burst", "Burst", "Burst", "Burst", 
"Burst", "Burst", "Medium.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming", "Medium.energy.swimming", "Medium.energy.swimming", 
"Medium.energy.swimming"), Roll_Avg = c(3.97084, 3.90604, 3.90738, 
3.80425, 3.4154, -0.993225, -0.940408, -0.55992, -0.791121, -1.83573, 
-3.41667, -14.0837, -14.9381, -16.4732, -16.6994, -15.5318, -18.2402, 
-19.4427, -22.8129, -27.009, -27.3907, 17.3778, 13.4861, 7.82564, 
4.63057, 6.94956, 14.3372, 22.0873, -11.5397, -11.7741, -11.4795, 
-10.7844, -10.5135, -11.0162, -90, -90, 11.0157, 6.13595, 2.2689, 
-0.710414, -5.56132, -12.0987, -9.70231, -7.13388, -5.41693, 
-4.23157, 2.11092, 2.19057, 1.5597, 0.637742, 1.17135, 3.41601, 
4.71664, 4.61525, -0.813111, -4.45238, -7.43746, -9.11626, -9.94338, 
-11.0361, -11.8852, -10.472, -5.12697, 2.61247, 9.80993, 17.307, 
10.5466, -4.01104, -7.40708, -2.72602, -5.43834, -5.22419, -4.8472, 
-4.43957, -1.67914, 2.39693, 7.84736, -9.7158, -8.70349, -8.22463, 
-8.22878, -9.43265, -0.527293, -0.283262, -0.614311, -0.380123, 
-0.344986, 7.73204, 7.47037, 7.00224, 7.01661, 7.38737, 7.83069, 
-1.83138, -1.7847, -1.68084, -1.61196, -1.49905, -1.61391, -1.46356, 
-0.986477, -0.806394, -0.883015, -0.840026, -0.727501, -1.15641, 
-1.28692, -1.38961, -1.43838, -1.42089, -1.27225)), class = "data.frame", row.names = c(NA, 
-111L))

问题根源 #

1. 排序失效：
   • 错误修改了behavssec$Behaviour而非目标数据集Seen2$Behaviour的因子水平
   • 指定的排序名称是空格分隔（如"Low energy swimming"），但数据中实际行为名称是点分隔（如"Low.energy.swimming"），两者不匹配导致排序规则失效
2. 颜色失效：
   • 自定义颜色的命名基于原始行为因子水平，未与修改后的排序levels对应
3. 缺失个体均值数据集：代码中使用了means但未提前计算，无法显示个体均值

修正后的完整代码 #

library(ggplot2)
library(dplyr)

# 1. 计算每个个体在各行为下的均值
means <- Seen2 %>%
  group_by(SharkID, Behaviour) %>%
  summarise(Roll_Avg = mean(Roll_Avg, na.rm = TRUE), .groups = "drop")

# 2. 定义行为排序顺序（与数据中的点分隔名称一致）
behav_order <- c("Burst", "High.energy.swimming", 
                 "Medium.energy.swimming", 
                 "Low.energy.swimming", "Travel",
                 "Ascending", "Descending")

# 3. 统一主数据集和均值数据集的因子水平，确保排序生效
Seen2$Behaviour <- factor(Seen2$Behaviour, levels = behav_order)
means$Behaviour <- factor(means$Behaviour, levels = behav_order)

# 4. 定义与排序匹配的自定义颜色
my_colors <- c("#CCFFFF", "#000000", "#7F7F7F", "#336699", "#008080", "#00CCFF", "#264AE2")
names(my_colors) <- behav_order
my_scale <- scale_fill_manual(name = "Behaviour", values = my_colors)

# 5. 绘制箱线图
ggplot(Seen2, aes(x = Behaviour, y = Roll_Avg, fill = Behaviour)) +
  geom_boxplot(outlier.shape = NA) +
  # 绘制个体均值抖动点，明确映射关系
  geom_point(data = means, aes(x = Behaviour, y = Roll_Avg), 
             size = 2, position = position_jitter(width = 0.2, height = 0.1)) +
  # 绘制整体均值红点
  stat_summary(fun = mean, geom = "point", shape = 20, size = 5, color = "red", fill = "red") +
  theme_classic() +
  my_scale +
  theme(axis.text.y = element_text(size = 16, angle = 0)) +
  ylim(-30, 30)

关键修改说明 #

• 用dplyr计算个体均值数据集means，确保每个个体-行为组合的均值正确
• 统一行为名称格式，同步修改主数据集和均值数据集的因子水平，保证排序规则生效
• 自定义颜色的命名直接使用排序后的行为列表，与因子levels完全匹配，解决颜色失效问题
• 明确指定抖动点的映射关系，避免数据集切换导致的映射错误

内容的提问来源于stack exchange，提问作者Sharklady