Reorder bars with ggplot with duplicate categories

Question

I'm trying to reorder the bars in my chart, some of which have duplicated category names, but some of them won't reorder. What am I doing wrong?

Sample data:
dyndom1 <- data.frame(Structure=c("T1R1vftd.web8", "6N4X.pdb", "T1R1vftd.web8", "T1R1vftd.web11", "6N52.pdb"),
                      Rotation_angle = c(31.4, 29, 4.9, 16.1, 28.1),
                      #Translation = c(0.5, 0.4, -0.3, 1.2),
                      Closure = c(99.9, 99.4, 98.9, 98.5, 96.7))
dyndom1

dyndom1$Structure <- fct_reorder(dyndom1$Structure, dyndom1$Closure)

ggplot(dyndom1, aes(fill=Structure, y=Closure, x=Structure)) + 
  geom_bar(position="dodge", stat="identity") +
  coord_flip()

Are there more than one observation per `Structure`? `fct_reorder` uses the median by default, you might want to use sum instead. — Jon Spring, Mar 12 '21 at 21:08
It sounds like the unusual thing you might be asking about is that you have multiple items in `Structure` which have the same name, but which represent distinct items and which you want to display as separate bars. e.g. you might have item/value pairs of A1, B2, C3, A4, and you want the y axis to show A B C A with those respective 1 2 3 4 values in the x axis. Is that right? — Jon Spring, Mar 12 '21 at 23:29
Yes, each structure item has a % closure, so it doesn't make sense to average or sum the closures. — glitterbox, Mar 13 '21 at 15:33

Jon Spring · Accepted Answer · 2021-03-13T18:41:37.093

0

EDIT: modified to answer clarified question

I think the hard part of the question here is how to show duplicated categories on the axis. I don't think there's a way to do this in ggplot2 without a hack. Here I've borrowed one from tidytext, which has a function called reorder_within which helps for sorting categories in different orders per facet. The hack is to make new categories with unique identifiers after a "___". By making the categories into an ordered factor and then removing the identifiers from the labels, I think we can get what you want.

library(dplyr)
dyndom1 %>%
  arrange(Closure) %>%
  mutate(Structure_unique = paste0(Structure, "___", row_number()) %>%
           forcats::fct_inorder()) %>%
  ggplot(aes(Closure, Structure_unique)) +
  geom_col() +
  tidytext::scale_y_reordered()

edited Mar 13 '21 at 18:41

answered Mar 12 '21 at 21:14

Jon Spring

55,165
4
35
53

Thank you for your answer, but it doesn't make sense to solve it like that for this data. I need to think more about it – glitterbox Mar 12 '21 at 21:30
So when you use `dyndom$Structure <- fct_reorder(dyndom$Structure, dyndom$Closure, .fun = sum)` that doesn't fix it? – Jon Spring Mar 12 '21 at 21:31
1

(It's also best practice on this site to include sample data in your question so that we can verify that it works instead of guessing about your data's characteristics.) – Jon Spring Mar 12 '21 at 21:34
Ah ok, I'm not sure how to add data... It doesn't fix it, but I realised that that approach wouldn't make sense with the data anyway – glitterbox Mar 12 '21 at 21:37
1

The standard guide for including data in your R question: https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example – Jon Spring Mar 12 '21 at 21:39
I've written a new answer that's hopefully more useful. – Jon Spring Mar 13 '21 at 18:43

Reorder bars with ggplot with duplicate categories

1 Answers1