Ordering bars in ggplot when dodged

Question

So I have a dataframe as follows:

structure(list(Reportable = c("Non", "Report", 
"Non", "Report", "Non", "Report", "Non", "Report", "Non", "Report"), Occupation = c("CONDUCTOR", 
"CONDUCTOR", "PASSENGER FROM THE DISTRICT NEEDING ASSISTANCE", "PASSENGER FROM THE DISTRICT NEEDING ASSISTANCE", 
"ENGINEER", "ENGINEER", "ASSISTANT", "ASSISTANT", 
"MEN", "MEN"
), count = c(88L, 173L, 42L, 60L, 24L, 48L, 21L, 25L, 19L, 15L
), pct = c("34%", "66%", "41%", "59%", "33%", "67%", "45.7%", 
"54.3%", "56%", "44%"), total = c("88 (34%)", "173 (66%)", "42 (41%)", 
"60 (59%)", "24 (33%)", "48 (67%)", "21 (45.7%)", "25 (54.3%)", 
"19 (56%)", "15 (44%)"), total1 = c("88\n(34%)", "173\n(66%)", 
"42\n(41%)", "60\n(59%)", "24\n(33%)", "48\n(67%)", "21\n(45.7%)", 
"25\n(54.3%)", "19\n(56%)", "15\n(44%)")), row.names = c(NA, 
-10L), class = c("tbl_df", "tbl", "data.frame"))

And I want to create a plot where these are ordered decreasingly by Reportable = "Report"

I can get close with the following but for whatever reason the reorder doesn't do what I need it to... I believe this is because there are two values for x.

library(tidyverse)
library(stringr)
ggplot(occ_counts, aes(fill=Reportable, y=count, x=as.factor(reorder(Occupation, -count)), label = total1)) +
  geom_bar(position="dodge", stat="identity")+
  aes(stringr::str_wrap(as.factor(Occupation), 15), count) +
  labs(x = "", y = "Injury Count", fill = "")+
  lims(y = c(0,230)) +
  geom_text(position = position_dodge(width = .9),     #move to center of bars
            vjust = -0.5,     #nudge above top of bar
            size = 3)

score 2 · Answer 1 · answered Aug 03 '20 at 08:49

you can use fct_reorder2 from the forcats package in tidyverse for that:

ggplot(df, aes(y=count, x= fct_reorder2(Occupation, Reportable, count) , 
               fill=Reportable,
               label = total1)) +
  geom_col(position = position_dodge(width = .9)) +
  geom_text(position = position_dodge(width = .9),     #move to center of bars
            vjust = -0.5,     #nudge above top of bar
            size = 3)  + 
labs(x = "", y = "Injury Count", fill = "")+
lims(y = c(0,230))  +
scale_x_discrete(labels = function(x) stringr::str_wrap(x, 15))

Ronak Shah · Accepted Answer · 2020-08-03T02:27:14.373

1

You can set your factor levels like this :

occ_counts$Occupation <- with(occ_counts, factor(Occupation, 
           levels = Occupation[Reportable == 'Report']
                    [order(-count[Reportable == 'Report'])]))

and then plot it as usual :

library(ggplot2)

ggplot(occ_counts, aes(fill=Reportable, y=count, x= Occupation, label = total1)) +
   geom_bar(position="dodge", stat="identity")+
   labs(x = "", y = "Injury Count", fill = "")+
   lims(y = c(0,230)) +
   geom_text(position = position_dodge(width = .9),    
             vjust = -0.5,    
             size = 3) +
  scale_x_discrete(labels = function(x) stringr::str_wrap(x, 15))

edited Aug 03 '20 at 02:27

answered Aug 03 '20 at 01:05

Ronak Shah

377,200
20
156
213

brilliant; but why doesnt reorder work in this case? – John Thomas Aug 03 '20 at 01:14
I don't use `reorder` much but I think the major issue is you need the values only for `Reportable = "Report"` which `reorder` code is not considering. – Ronak Shah Aug 03 '20 at 01:15
It defaults to alphabetical – John Thomas Aug 03 '20 at 01:52
I have removed `aes(stringr::str_wrap(as.factor(Occupation), 15), count)` from your code in my answer. Are you sure you don't have that in your code as well? – Ronak Shah Aug 03 '20 at 02:03
Yes, that is critical to be included, since some bars in the full data is very long – John Thomas Aug 03 '20 at 02:07
Yeah to summarize, I need to be able to plot using the lines in my ggplot above. the stringr line is required here so how can I order these given that condition? – John Thomas Aug 03 '20 at 02:12
Updated sample to reflect – John Thomas Aug 03 '20 at 02:17
You can do it via `scale_x_discrete`. See updated answer. – Ronak Shah Aug 03 '20 at 02:27

Ordering bars in ggplot when dodged

2 Answers2

Linked