I am searching for a way to calculate and plot significance of certain levels within a grouped barplot in R using ggplot and ggpubr. There is some information available, but non of them seems to work for that special issue. The base data appear to be an aggregation table which looks like that:
survey1 <- as.data.frame(replicate(3,sample(0:1,4,rep=TRUE)))
survey1$V1 <- runif(4)
survey1$V1 <- survey1$V1/sum(survey1$V1)
survey1$V2 <- "survey1"
survey1$V3 <- 1:4
survey1$variable <- "F2500"
colnames(survey1)[1] <- "percent"
colnames(survey1)[2] <- "survey"
colnames(survey1)[3] <- "level"
survey2 <- as.data.frame(replicate(3,sample(0:1,4,rep=TRUE)))
survey2$V1 <- runif(4)
survey2$V1 <- survey2$V1/sum(survey2$V1)
survey2$V2 <- "survey2"
survey2$V3 <- 1:4
survey2$variable <- "F2500"
colnames(survey2)[1] <- "percent"
colnames(survey2)[2] <- "survey"
colnames(survey2)[3] <- "level"
sum.tab <- rbind(survey1, survey2)
My approach to plot this variable was the following:
library(ggpubr)
library(ggplot2)
ggbarplot(sum.tab, x = "level", y = "percent", fill = "survey", color = NA,
position = position_dodge2(preserve = "single", padding = 0)) +
stat_compare_means(aes(group = survey), label = "p.format")
However, it shows a result, but it doesn't seem right. I am searching for an approach to run several dozen variables with a similar plot code within a function...but I am not right there, yet.
Sorry, if I have overseen prior posts regarding that issue (maybe I didn't understand them enough).
Thanks for your help!
question from:https://stackoverflow.com/questions/65849029/r-show-level-significance-in-grouped-barplot
