ggplot 2：两个分组变量的均值和95％CI的点图

Question

我正在尝试生成一个点图并为两个独立的分组变量添加平均值和95％CI（因此我总共有2 x 2个组）。

这是我的数据：

GluNorm Gender  Treatment
1.317   Male    NAC
1.278   Male    SAL
1.302   Male    SAL
1.376   Male    NAC
1.279   Male    NAC
1.308   Male    SAL
1.451   Male    NAC
1.244   Male    NAC
1.411   Male    SAL
1.16    Male    NAC
1.159   Male    NAC
1.42    Male    SAL
1.407   Male    SAL
1.62    Male    SAL
1.167   Male    SAL
1.377   Male    NAC
1.393   Female  SAL
1.203   Female  NAC
1.191   Female  NAC
1.132   Female  SAL
1.191   Female  SAL
1.589   Female  SAL
1.169   Female  NAC
1.155   Female  SAL
1.249   Female  NAC
1.401   Female  NAC
1.455   Female  SAL
1.481   Female  SAL
1.293   Female  NAC
1.332   Female  NAC
1.462   Female  SAL

这就是我想要的情节：

我做了以下尝试：

g<- ggplot(Data, aes(x = Gender, y = GluNorm, fill=Group))+
  geom_dotplot(binaxis='y', stackdir='center',stackratio=1.2, dotsize=1.2, binwidth=0.02, position=position_dodge(0.8)) 


#function that outputs mean, lower limit and upper limit of 95% CI
data_summary <- function(x) {
  m <- mean(x, na.rm=TRUE)
  sem <-sd(x, na.rm=TRUE)/sqrt(sum(!is.na(x)))
  ymin<-m-1.96*sem
  ymax<-m+1.96*sem
  return(c(y=m,ymin=ymin,ymax=ymax))
}

g +  stat_summary(fun.data=data_summary, color="red")

得到以下结果：

有没有人可以建议如何为每个性别x治疗组获得单独的点和条？

谢谢！

Answer 1

如果您以不同的数量躲避多重点和置信区间，图表将更容易阅读，以便不重叠：

pd1 = position_dodge(0.2)
pd2 = position_dodge(0.65)

ggplot(Data, aes(x = Gender, y = GluNorm, fill=Treatment, color=Treatment))+
  geom_dotplot(binaxis='y', stackdir='center',stackratio=1.2, 
               dotsize=0.8, binwidth=0.02, position=pd2) +
  stat_summary(fun.data=data_summary, position=pd1, geom="errorbar", width=0.05) +
  stat_summary(fun.data=data_summary, position=pd1, geom="point", size=2) +
  scale_fill_manual(values=hcl(c(15,195), 100, 60)) +
  scale_color_manual(values=hcl(c(15,195), 50, 40)) +
  theme_bw()

您也可以使用内置函数代替data_summary函数。下面我们使用bootstrapping获得95％的CI：

ggplot(Data, aes(x = Gender, y = GluNorm, fill=Treatment, color=Treatment))+
  geom_dotplot(binaxis='y', stackdir='center',stackratio=1.2, 
               dotsize=0.8, binwidth=0.02, position=pd2) +
  stat_summary(fun.data=mean_cl_boot, position=pd1, geom="errorbar", width=0.05) +
  stat_summary(fun.y=mean, position=pd1, geom="point", size=2) +
  scale_fill_manual(values=hcl(c(15,195), 100, 60)) +
  scale_color_manual(values=hcl(c(15,195), 50, 40)) +
  theme_bw()

您可以直接绘制点，而不是点图，添加抖动以避免重叠：

ggplot(Data, aes(x = Gender, y = GluNorm, fill=Treatment, color=Treatment))+
  geom_point(position=position_jitterdodge(dodge.width=0.65, jitter.height=0, jitter.width=0.25),
             alpha=0.7) +
  stat_summary(fun.data=mean_cl_boot, geom="errorbar", width=0.05, position=pd1) +
  stat_summary(fun.y=mean, geom="point", size=2, position=pd1) +
  scale_fill_manual(values=hcl(c(15,195), 100, 60)) +
  scale_color_manual(values=hcl(c(15,195), 50, 40)) +
  theme_bw()

或者也许是一个箱子图：

pd1 = position_dodge(0.4)

ggplot(Data, aes(x = Gender, y = GluNorm, fill=Treatment, color=Treatment))+
  geom_boxplot(position=pd1, width=0.3, alpha=0.2, color="#00000080", 
               lwd=0.4, fatten=1.5) +
  stat_summary(fun.data=mean_cl_boot, geom="errorbar", width=0.1, position=pd1, size=0.7) +
  stat_summary(fun.y=mean, geom="point", size=2, position=pd1) +
  scale_fill_manual(values=hcl(c(15,195), 100, 80)) +
  scale_color_manual(values=hcl(c(15,195), 50, 30)) +
  theme_bw()