如何在ggplot中合并两组带置信区间的分组散点误差棒图

要把两组Bootstrap结果的图表合并，最简洁的方式是先将两个数据框合并并添加分组标识，再通过ggplot的分组或分面功能实现。以下是具体实现步骤：

步骤1：合并数据框 #

给d1和d2分别添加一个用于区分数据集的列（比如dataset），然后合并成一个数据框：

# 添加分组标识
d1$dataset <- "d1"
d2$dataset <- "d2"

# 合并数据
combined_data <- rbind(d1, d2)

步骤2：绘制合并后的图表 #

方式1：同一面板内区分两组数据（错位显示） #

通过position_dodge让同一pqual下的两组数据错开显示，同时用颜色区分数据集：

ggplot(data = combined_data, aes(x = pqual, y = V1, color = dataset, group = dataset)) +
  geom_point(position = position_dodge(width = 0.3), size = 3) +
  geom_errorbar(aes(ymin = lower, ymax = upper), 
                width = 0.2, position = position_dodge(width = 0.3)) +
  scale_y_continuous(limits = c(0.5, 2), breaks = c(0.5, 1, 1.5, 2)) +
  scale_colour_manual(values = c("d1" = "aquamarine3", "d2" = "darkgoldenrod3"),
                      labels = c("Dataset 1", "Dataset 2")) +
  theme_grey(base_size = 22) +
  labs(title = "Q10 Bootstrap Results", x = "pqual", y = "Q10", color = "Dataset")

方式2：分面显示（独立栏位） #

如果希望两组数据保持独立但在同一图表中，用facet_wrap实现横向/纵向分栏：

ggplot(data = combined_data, aes(x = pqual, y = V1, color = pqual)) +
  geom_point(size = 3) +
  geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.2) +
  scale_y_continuous(limits = c(0.5, 2), breaks = c(0.5, 1, 1.5, 2)) +
  scale_colour_manual(values = c("scst" = "aquamarine3",
                                 "svar" = "aquamarine4",
                                 "ycst" = "darkgoldenrod3",
                                 "yvar" = "darkgoldenrod4")) +
  theme_grey(base_size = 22) +
  labs(title = "Q10 Bootstrap Results", x = "pqual", y = "Q10") +
  facet_wrap(~dataset)  # 横向分栏，若需纵向可改用facet_grid(dataset~.)

原始数据与参考代码 #

你的原始数据和绘图代码如下：

# 原始数据
d1 <- structure(list(pqual = c("scst", "svar", "ycst", "yvar"), V1 = c(1.57018534500418, 
0.813141641335401, 1.83263172676745, 1.08637937362865), lower = c(1.46430733516951, 
0.754261886214852, 1.83263172676745, 1.04296652726617), median = c(1.56519136055835, 
0.813147324264586, 1.83263172676745, 1.08589629982024), upper = c(1.684225201556, 
0.875420698794502, 1.83263172676745, 1.12960240657212)), class = "data.frame", row.names = c("bootscstd", 
"bootsvard", "bootycstd", "bootyvard"))

d2 <- structure(list(pqual = c("scst", "svar", "ycst", "yvar"), V1 = c(1.52392776239251, 
0.795105955855431, 1.7049070356902, 0.866155843608828), lower = c(1.42095510456812, 
0.740169509961879, 1.67591486183307, 0.831244001139425), median = c(1.52198888072917, 
0.794298998427312, 1.70461432040644, 0.864517118120389), upper = c(1.63130199180125, 
0.85303654089534, 1.73294982674048, 0.908100088043788)), class = "data.frame", row.names = c("bootscst", 
"bootsvar", "bootycst", "bootyvar"))

# 原始绘图代码
Q10dplot <- ggplot(data=Q10d, aes(x=pqual, y=V1, col=pqual))+
  geom_point()+
  geom_errorbar(aes(ymin=lower, ymax=upper), width=.2)+
  scale_y_continuous(limits=c(0.5,2), breaks = c(0.5,1,1.5,2))+
  scale_colour_manual(values=c("scst" = "aquamarine3",
                               "svar" = "aquamarine4",
                               "ycst" = "darkgoldenrod3",
                               "yvar" = "darkgoldenrod4"))+
  theme_minimal()+
  ggtitle("Q10 d")+
  xlab("pqual") + ylab("Q10")+
  theme_grey(base_size = 22)
Q10dplot

内容的提问来源于stack exchange，提问作者Nate Trf