How to reorder plots in combined ggplot2 graph?

Question

I have seen the solutions to reordering subplots when it's just one object being plotted (e.g. mydata), but I am not sure how to do this when there are multiple objects being plotted (in this instance, mydata1 and mydata2). I would like to switch the order of the violins such that Treatment2 is on the left, and Treatment1 is on the right, instead of vice-versa like I currently have it:

mycp <- ggplot() + geom_violin(data = mydata1, aes(x= treatment, y = Myc_List1, fill = Myc_List1, colour="Myc Pathway (Treatment1)")) + 
  geom_violin(data = mydata2, aes(x= treatment, y = Myc_List1, fill = Myc_List1, colour = "Myc Pathway (Treatment2)"))

When I try solutions such as in Ordering of bars in ggplot, or the following solution posed at https://www.r-graph-gallery.com/22-order-boxplot-labels-by-names.html, this graph remains unchanged.

Hopefully this makes sense, and thank you for reading!

UPDATE

Here is another solution as well from https://www.datanovia.com/en/blog/how-to-change-ggplot-legend-order/

mydata$treatment<- factor(mydata$treatment, levels = c("Treatment2", "Treatment1"))

Without access to `mydata1` we are just guessing. – IRTFM Jun 09 '21 at 00:44 — IRTFM, Jun 09 '21 at 00:44

score 2 · Accepted Answer · answered Jun 08 '21 at 23:52

2

I'm not sure how to reorder factors in this case, but you can change the x axis scale to get the desired result, e.g.

library(tidyverse)

data("Puromycin")
dat1 <- Puromycin %>% 
  filter(state == "treated")
dat2 <- Puromycin %>% 
  filter(state == "untreated")

mycp <- ggplot() +
  geom_violin(data = dat1, aes(x= state, y = conc, colour = "Puromycin (Treatment1)")) + 
  geom_violin(data = dat2, aes(x= state, y = conc, colour = "Puromycin (Treatment2)")) 
mycp

mycp2 <- ggplot() +
  geom_violin(data = dat1, aes(x = state, y = conc, colour = "Puromycin (Treatment1)")) +
  geom_violin(data = dat2, aes(x = state, y = conc, colour = "Puromycin (Treatment2)")) +
  scale_x_discrete(limits = c("untreated", "treated"))
mycp2

answered Jun 08 '21 at 23:52

jared_mamrot

22,354
4
21
46

The first part of your first example and then continuing with the second example worked for me @jared_mamrot! Thank you so much! `library(tidyverse) data("Puromycin") dat1 <- Puromycin %>% filter(state == "treated") dat2 <- Puromycin %>% filter(state == "untreated") mycp2 <- ggplot() + geom_violin(data = dat1, aes(x = state, y = conc, colour = "Puromycin (Treatment1)")) + geom_violin(data = dat2, aes(x = state, y = conc, colour = "Puromycin (Treatment2)")) + scale_x_discrete(limits = c("untreated", "treated")) mycp2` – KaitS Jun 09 '21 at 04:40
1

Yes @KaitS, my answer is an example of a [minimal reproducible example](https://stackoverflow.com/help/minimal-reproducible-example) (MRE). The first part of the answer provides a publicly-available dataset and the minimal amount of code required to replicate the problem (i.e. the "starting point"), and the second part shows a potential solution. On StackOverflow you are strongly encouraged to create an MRE when you ask a question as it helps us troubleshoot the issue (please use an MRE in future questions, including your own attempts to solve the issue). I'm glad it solved your problem :) – jared_mamrot Jun 09 '21 at 04:46

eipi10 · Answer 2 · 2021-06-09T00:05:58.090

Stack the data into a single data frame and set the order by converting treatment to a factor. In your example, the colors and legend are redundant, since you can label the x-axis values to describe each treatment, or change the x-axis title to "Myc Pathway", but the code below in any case shows how to get the ordering.

library(tidyverse)

bind_rows(mydata1, mydata2) %>%
  mutate(treatment = factor(treatment, levels=paste0("Treatment", c(2,1)) %>%
  ggplot(aes(treatment, Myc_List1, colour=treatment)) +
    geom_violin()

Here's a reproducible example:

library(tidyverse)
theme_set(theme_bw(base_size=15))

# Create two separate data frames to start with
d1=iris %>% filter(Species=="setosa")
d2=iris %>% filter(Species=="versicolor") 

bind_rows(d1, d2) %>% 
  mutate(Species = factor(Species, levels=c("versicolor", "setosa"))) %>% 
  ggplot(aes(Species, Petal.Width, colour=Species)) +
    geom_violin()

Following your second example @eipi10, unfortunately I got this error: "Error: ` Problem with mutate()` column `treatment`. ? `treatment = +...`. x `data` must be a data frame, or other object coercible by `fortify()`, not a factor" . Thank you for your help though! I really do appreciate it! — KaitS, Jun 09 '21 at 04:26

How to reorder plots in combined ggplot2 graph?

UPDATE

2 Answers2

Linked