ggplot statistical differences in plot labels (reproducible code included) - r

Question

I have a code that generates two plots (actually from different datasets) like this one:
#Plot 1
p1 <- ggplot(mtcars,aes(x=factor(cyl),fill=factor(gear)))+
geom_bar(position="fill")+
geom_text(aes(label=scales::percent(..count../sum(..count..))),
stat='count',position=position_fill(vjust=0.5))
#Plot 2
p2 <- ggplot(mtcars,aes(x=factor(cyl),fill=factor(gear)))+
geom_bar(position="fill")+
geom_text(aes(label=scales::percent(..count../sum(..count..))),
stat='count',position=position_fill(vjust=0.5))
plot <- p1 + p2
plot
Is it possible using gglpot o other library to test statistical differences among factos and if there are statistical differences among them to change the label from 25% from something like "25% ↑" or "25% ** " so what I want is to compare values and change labeling to include statistical differences. In my example values are the same but in reality plots are coming from different datasets.

Answer 1

As MrFlick mentioned, ggplot might not be the right tool to do the calculations. But once you have your calculations, you could do something like that
# some date with calculated levels of significance
dplyr::tibble(YEAR=rep(c(2019,2020),eac=3),
GRP=rep(c("A","B","C"),2),
VAL=c(20,100,30,25,70,30),
SIG=rep(c("*","***",""),2)) %>%
# create labels
dplyr::group_by(GRP) %>%
dplyr::mutate(LABEL=dplyr::case_when(VAL/sum(VAL)<0.5 ~ paste("<",SIG),
VAL/sum(VAL)>0.5 ~ paste(">",SIG),
TRUE ~ paste(""))) %>%
dplyr::ungroup() %>%
# calculate percentages
dplyr::group_by(YEAR) %>%
dplyr::mutate(VAL=VAL/sum(VAL)) %>%
dplyr::ungroup() %>%
# plot data: combining percentages and sig-levels as label
ggplot2::ggplot(ggplot2::aes(x=YEAR,
y=VAL,
fill=GRP,
label=glue::glue("{scales::percent(VAL)} {LABEL}"))) +
ggplot2::geom_bar(stat="identity") +
ggplot2::geom_text(position=ggplot2::position_fill(0.5))