How can I set median crossbars to align within factors?

Question

I have a data frame like so:

my_df <- structure(list(SampleID = c("sample01", "sample02", "sample03", 
"sample04", "sample05", "sample06", "sample07", "sample08", "sample09", 
"sample10", "sample11", "sample12", "sample13", "sample14", "sample15", 
"sample16", "sample17", "sample18", "sample19", "sample20"), 
    y = c(1.68547922357333, 0.717650914301956, 1.18156420566867, 
    1.31643130248052, 1.2021341615705, 0.946937741954258, 1.75576099871947, 
    0.952670480793451, 2.00921185693852, 0.968642950473789, 1.65243482711174, 
    2.14332269635055, 0.30556964944383, 0.860605616591314, 0.933339331803171, 
    1.31797519903504, 0.857873539291964, -0.328227710452388, 
    -0.22023346428776, 1.6600566728651), week = structure(c(1L, 
    2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 3L, 1L, 2L, 
    3L, 1L, 2L, 3L), .Label = c("0", "3", "6"), class = "factor"), 
    grumpy = structure(c(2L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 
    2L, 2L, 2L, 1L, 1L, 2L, 2L, 2L, 1L, 1L, 1L), .Label = c("No", 
    "Yes"), class = "factor"), week_grumpy = structure(c(2L, 
    4L, 6L, 2L, 4L, 6L, 1L, 3L, 5L, 2L, 4L, 6L, 1L, 5L, 2L, 4L, 
    6L, 1L, 3L, 5L), .Label = c("0 No", "0 Yes", "3 No", "3 Yes", 
    "6 No", "6 Yes"), class = "factor")), class = c("spec_tbl_df", 
"tbl_df", "tbl", "data.frame"), row.names = c(NA, -20L))

#packages needed if you don't have
install.packages("ggbeeswarm")
install.packages("ggplot2")

This is typically how I graph:

library(ggplot2)
library(ggbeeswarm)

ggplot(data = my_df, aes(x=week, y=y, color=grumpy)) +
    geom_quasirandom(dodge.width = 0.75)

Which is nice because it separates the colors rather nicely. Nowadays, I like to add a median crossbars to further show the differences between groups. Like so:

ggplot(data = my_df, aes(x=week, y=y, color=grumpy)) +
    geom_quasirandom(dodge.width = 0.75) +
    stat_summary(aes(group = grumpy), fun = median, fun.min = median, fun.max = median, geom = "crossbar", color = "black", width = 0.7, lwd = 0.2)

Now, what I would love to have is the median crossbars to align with the colors within each factor on the x-axis. Is there a way to do this within R? Or am I relegated to manually editing the crossbars to line up?

Here's is one thing I have tried:

ggplot(data = my_df, aes(x=week_grumpy, y=y, color=grumpy)) +
    geom_jitter(width = 0.1) +
    stat_summary(aes(group = grumpy), fun = median, fun.min = median, fun.max = median, geom = "crossbar", color = "black", width = 0.7, lwd = 0.2)

But now the x-axis is not the way I want it (However, it would be easier to manually edit in something like Inkscape than the previous example).

I've found some hints here and here but have yet to arrive at a satisfactory solution.

score 2 · Accepted Answer · answered Aug 31 '20 at 16:54

What you are looking for is to dodge the crossbar geom. For example:

ggplot(data = my_df, aes(x=week, y=y, color=grumpy)) +
  geom_quasirandom(dodge.width = 0.75) +
  stat_summary(
    aes(group = grumpy), fun = median, fun.min = median, fun.max = median,
    geom = "crossbar", color = "black", width = 0.7, lwd = 0.2,

    # add this bit here to your stat_summary function
    position=position_dodge(width=0.75)
  )

It seems that geom_quasirandom() is acting here very similarly to geom_point(position=position_jitterdodge(dodge.width=0.75)). In this case, since dodge.width is specified in geom_quasirandom(), you use the same width for position_dodge in the crossbar geom.

Note: you may want to play around with aesthetic formatting to be able to make the distinction a bit more clear what the crossbars are telling you, but this should answer your question.

How can I set median crossbars to align within factors?

1 Answers1