使用facet_grid在geom_boxplot上的Tukey测试结果 [英] Tukey test results on geom_boxplot with facet_grid

问题描述

我希望将代表Tukey HSD结果的字母放在我在ggplot中制作的箱图中.我知道其他帖子(在ggplot boxplot上使用Tukeys post-hoc ， Tukey在R中箱图上的结果) 但就我而言，我与facet_grid合作，但不确定如何继续.当我创建许多箱形图时，我很想自动化一些东西.任何帮助我入门的提示将不胜感激.非常感谢

I'd love to place letters representing the results of a Tukey's HSD on boxplots I make in ggplot. I am aware of other posts (Tukeys post-hoc on ggplot boxplot, Tukey's results on boxplot in R) but in my case, I work with facet_grid and I am not sure how to proceed. As I am creating a lot of boxplots, I would love to have something automated. Any tips to help me get me started would be greatly appreciated. Thanks a lot

library(multcompView)
library(ggplot2)
#sample dataframe
df <- data.frame(var_facet=c('facet1', 'facet1', 'facet1', 'facet1', 'facet1', 
                             'facet1', 'facet1', 'facet1', 'facet1', 'facet1', 
                             'facet2', 'facet2', 'facet2', 'facet2', 'facet2', 
                             'facet2', 'facet2', 'facet2', 'facet2', 'facet2'), 
                 class_x=c("Q1", "Q1", "Q1", "Q2", "Q2", "Q2", "Q3", "Q3", "Q3", "Q3", 
                              "Q1", "Q1", "Q1", "Q2", "Q2", "Q2", "Q3", "Q3", "Q3", "Q3"),
                 value_y=c(12, 21, 35, 42, 56, 1, 4, 3, 4, 5, 2, 2, 2, 
                           2, 2, 4, 4, 4, 4, 4))

#function 
generate_label_df <- function(TUKEY, variable){

  # Extract labels and factor levels from Tukey post-hoc 
  Tukey.levels <- variable[,4]
  Tukey.labels <- data.frame(multcompLetters(Tukey.levels)['Letters'])

  #I need to put the labels in the same order as in the boxplot :
  Tukey.labels$treatment=rownames(Tukey.labels)
  Tukey.labels=Tukey.labels[order(Tukey.labels$treatment) , ]
  return(Tukey.labels)
}

model=lm(value_y~class_x, data=df)
ANOVA=aov(model)
TUKEY <- TukeyHSD(ANOVA)
plot(TUKEY , las=1 , col="brown" )

labels <- generate_label_df(TUKEY , TUKEY$`class_x`)#generate labels using function

names(labels) <- c('Letters','class_x')#rename columns for merging

yvalue <- aggregate(.~class_x, data=df, mean)# obtain letter position for y axis using means

final <- merge(labels,yvalue) #merge dataframes

#ggplot with facet_grid
p <- ggplot(data=df, aes(class_x , value_y, fill=class_x))+  
  geom_boxplot(outlier.shape = NA) +
  facet_grid(~var_facet)
p <- p +  scale_y_continuous(limits = quantile(df$value_y, c(0.1, 0.9), na.rm=T))
p <- p + theme_minimal()
p <- p + scale_fill_brewer(palette="Greens")
p <- p + theme(legend.position="none")
p <- p + theme(axis.text.x=element_text(angle=45, hjust=1))
p <- p + geom_text(data = final, 
                   aes(x = class_x, y = value_y, label = Letters), 
                   vjust=-3.5,hjust=-.5)
p

推荐答案

也许这就是您想要的.

p <- ggplot(data=df, aes(x=class_x , y=value_y, fill=class_x)) +  
     geom_boxplot(outlier.shape=NA) +
     facet_grid(~var_facet) +
     scale_fill_brewer(palette="Greens") +
     theme_minimal() +
     theme(legend.position="none") +
     theme(axis.text.x=element_text(angle=45, hjust=1)) 

for (facetk in as.character(unique(df$var_facet))) {   
  subdf <- subset(df, var_facet==facetk)
  model=lm(value_y ~ class_x, data=subdf)
  ANOVA=aov(model)
  TUKEY <- TukeyHSD(ANOVA)

  labels <- generate_label_df(TUKEY , TUKEY$`class_x`)
  names(labels) <- c('Letters','class_x')
  yvalue <- aggregate(.~class_x, data=subdf, quantile, probs=.75)  
  final <- merge(labels, yvalue)
  final$var_facet <-  facetk

  p <- p + geom_text(data = final,  aes(x=class_x, y=value_y, label=Letters), 
                   vjust=-1.5, hjust=-.5)
}
p

这篇关于使用facet_grid在geom_boxplot上的Tukey测试结果的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！