Ggplot - 如何添加 2 个不同的误差线？

Question

我尝试制作具有 2 个不同置信区间的条形图 - 一个代表样本中女性的比例，另一个代表男性的比例。每个类别当然有两个不同的置信区间，我怎样才能制作只有 2 个而不是 4 个置信区间的图表？

  ggplot(data, aes(x= GENDER)) + 
  geom_bar(aes(y = (..count..)/sum(..count..)), stat="count", fill=c("deeppink","deepskyblue"), alpha=0.7) +
  scale_y_continuous("Percent",labels = scales::percent)+
  geom_text(aes(label = scales::percent((..count..)/sum(..count..)),
                y= ((..count..)/sum(..count..))), stat="count",vjust = 5) +
  geom_errorbar (aes(ymin = ymin1, ymax =ymax1), width=0.4, colour = "red", alpha =0.9, size= 1.3)+
  geom_errorbar (aes(ymin = ymin2, ymax =ymax2), width=0.4, colour = "red", alpha =0.9, size= 1.3)

先谢谢了！

Answer 1

您只需要调用一次 geom_errorbar，其中 ymin 和 ymax 是向量。按照您的编码方式，ggplot 在两个位置绘制两个误差线，因为它期望 ymin 和 ymax 的位置向量等于图中的条数。例如

ymin = c(ymin1,ymin2)
ymax = c(ymax1,ymax2)
ggplot(data, aes(x= GENDER)) + 
  geom_bar(aes(y = (..count..)/sum(..count..)), stat="count", fill=c("deeppink","deepskyblue"), alpha=0.7) +
  scale_y_continuous("Percent",labels = scales::percent)+
  geom_text(aes(label = scales::percent((..count..)/sum(..count..)),
                y= ((..count..)/sum(..count..))), stat="count",vjust = 5) +
  geom_errorbar (aes(ymin = ymin, ymax =ymax), width=0.4, colour = "red", alpha =0.9, size= 1.3)

Answer 2

您还可以使用 dplyr::summarise() 提前管理数据，以创建一个数据集，该数据集每组只有一个观察值，具有您想要的任何汇总值。这使得绘图代码更加精简。我在carData 包中的Chile 数据上执行此操作

data("Chile", package="carData")
library(dplyr)
library(ggplot2)
Chile %>% 
  group_by(sex) %>% 
  dplyr::summarise(n_g= n(), 
            n = sum(!is.na(Chile$sex)),
            prop = n_g/n, 
            ymin = binom.test(n_g, n)$conf.int[1], 
            ymax = binom.test(n_g, n)$conf.int[2]) %>% 
  ggplot(aes(x=sex, y=prop, ymin=ymin, ymax=ymax) )+ 
    geom_bar(stat="identity", alpha=.7, fill=c("deeppink","deepskyblue")) + 
    geom_errorbar(width=.4, colour="red", alpha=.9, size=1.3) + 
    geom_text(aes(label = scales::percent(prop)), vjust = 5) 

---

编辑

更改为在 OP 的评论中使用 prop.test()。

data("Chile", package="carData")
 library(dplyr)
 library(ggplot2)
 data <- Chile
 data <- data %>% rename("GENDER" = "sex")
 
 data %>% 
   group_by(GENDER) %>% 
   dplyr::summarise(n_g= n(), 
                    n = sum(!is.na(data$GENDER)),
                    prop = n_g/n,  
                    ymin = prop.test(n_g, n, p=.7, alternative="greater")$conf.int[1], 
                    ymax = prop.test(n_g, n, p=.7, alternative="greater")$conf.int[2]) %>% 
   ggplot(aes(x=GENDER, y=prop, ymin=ymin, ymax=ymax) )+ 
   geom_bar(stat="identity", alpha=.7, fill=c("deeppink","deepskyblue")) + 
   geom_errorbar(width=.4, colour="red", alpha=.9, size=1.3) + 
   geom_text(aes(label = scales::percent(prop)), vjust = 5)