Standard column width in facetted and grouped ggplot bar plot - r

Question

I've made a bar chart using ggplot with grouped data, and facetted with facet_grid. The column widths are inconsistent, so I want to make them all the same. I've read this can be done with preserve="single, but it seems to mess up the position dodging. Any idea how to prevent this happening??
Here is a small sample of the data:
data <- structure(list(grp2 = structure(c(3L, 3L, 3L, 3L, 3L, 3L, 3L,
7L, 7L, 7L, 7L, 7L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 7L, 7L, 7L, 7L,
7L), .Label = c("CSF1", "CSF2", "PC", "NC", "GPC", "GNC", "standard"
), class = "factor"), label2 = structure(c(7L, 8L, 9L, 7L, 8L,
9L, 7L, 15L, 15L, 15L, 15L, 15L, 7L, 8L, 9L, 7L, 8L, 9L, 7L,
15L, 15L, 15L, 15L, 15L), .Label = c("CSF1_raw", "CSF1_supernatant",
"CSF1_pellet", "CSF2_raw", "CSF2_supernatant", "CSF2_pellet",
"PC_raw", "PC_supernatant", "PC_pellet", "NC_raw", "NC_supernatant",
"NC_pellet", "GPC", "GNC", "standard", "NC"), class = "factor"),
mda_label = structure(c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 1L, 1L, 1L, 1L, 1L
), .Label = c("none", "mda_20", "mda_200"), class = "factor"),
conc = c(`7` = 0, `8` = 0, `9` = 0.324886127298521, `55` = 4.14765656994934,
`56` = 1.16840050032707, `57` = 8.33529714053568, `76` = 10.6220645144775,
`77` = 48.9241552191721, `78` = 4.51513315624087, `79` = 1.03887911533275,
`80` = 0.0445944796011582, `81` = 0.00484116548901831, `89` = 0,
`90` = 0, `91` = 0.322922569348207, `137` = 6.38488684568018,
`138` = 1.68909814271646, `139` = 7.61828609738757, `158` = 15.3082130743032,
`159` = 41.3127531345335, `160` = 4.64193087683391, `161` = 0.411672491030815,
`162` = 0.0568193835425769, `163` = 0.00439419098560105)), row.names = c(NA,
-24L), class = c("tbl_df", "tbl", "data.frame"))
Here's the initial plot:
ggplot(data, aes(x=label2, y=conc, colour=mda_label, fill=mda_label)) +
facet_grid(. ~ grp2, scales="free_x", space="free") +
stat_summary(fun = mean, geom = "bar", position = position_dodge()) +
stat_summary(fun.data = mean_se, geom = "errorbar", colour="black", width=0.5,
position = position_dodge(width=0.9)) +
geom_point(position = position_dodge(width=0.9), pch=21, colour="black") +
scale_y_continuous(trans='pseudo_log',
labels = scales::number_format(accuracy=0.01)) +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
But when I try to standardise the column widths with preserve="single", it gets messed up:
ggplot(data, aes(x=label2, y=conc, colour=mda_label, fill=mda_label)) +
facet_grid(. ~ grp2, scales="free_x", space="free") +
stat_summary(fun = mean, geom = "bar", position = position_dodge(preserve="single")) +
stat_summary(fun.data = mean_se, geom = "errorbar", colour="black", width=0.5,
position = position_dodge(width=0.9, preserve="single")) +
geom_point(position = position_dodge(width=0.9, preserve="single"), pch=21, colour="black") +
scale_y_continuous(trans='pseudo_log',
labels = scales::number_format(accuracy=0.01)) +
theme(axis.text.x = element_text(angle = 45, hjust = 1))

Answer 1

Since you're using data that as 0 values, you could make the 0 values for the other 'mda_label' on grp2/label2 standard categories.
data <- rbind(data, data.frame(grp2 = c("standard", "standard"),
label2 = c("standard", "standard"),
mda_label = c("mda_20", "mda_200"),
conc = c(0, 0)))
Also you never actually make the bar plot
data %>%
ggplot(aes(label2, conc, fill = mda_label)) +
geom_col(position = position_dodge(width = 1)) +
facet_grid(. ~ grp2, scales = "free", space = "free")