CODEWITHSUNDEEP
rggplot2plot
scoste
scoste

Reputation: 21

"fill" argument creates overlapping points in geom_dotplot in R

I am using geom_dotplot and want to differentiate by color which points are in one group versus another. I have successfully done this by adding "fill=group" to the aes() in geom_dotplot().

Here is some sample code that reproduces the error:

set.seed(124)
df <- data.frame(Group = rep(c("control","treatment"),20), Response = sample(1:10,40, replace = T), Recovered = rep(c("no","no","no","no","yes"),4))
ggplot() + geom_dotplot(data = df, aes(x = Group, y = Response, fill = Recovered),binaxis = "y", stackdir = "center", alpha = 0.3) + coord_flip()

However, now the package no longer stacks the points in one group alongside the other, but instead overlaps them, hiding some of the data.

I can set alpha = 0.5 to see where this overlap occurs, but I would much rather have it plot all the points alongside each other and simply color some of the points. Does anyone know how to do this?

I know I can set position_dodge to a small amount, but would prefer not to as that messes up axis interpretation.

Edit: output of dput(df) is:

structure(list(Group = structure(c(1L, 2L, 1L, 2L, 1L, 2L, 1L, 
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 
2L), .Label = c("control", "treatment"), class = "factor"), Response = c(1L, 
5L, 6L, 4L, 3L, 3L, 6L, 5L, 10L, 3L, 8L, 9L, 8L, 9L, 5L, 1L, 
6L, 8L, 9L, 1L, 7L, 7L, 1L, 5L, 4L, 3L, 9L, 3L, 9L, 4L, 9L, 4L, 
5L, 9L, 2L, 10L, 2L, 10L, 2L, 4L), Recovered = structure(c(1L, 
1L, 1L, 1L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 
1L, 1L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 1L, 
1L, 2L, 1L, 1L, 1L, 1L, 2L), class = "factor", .Label = c("no", 
"yes"))), .Names = c("Group", "Response", "Recovered"), row.names = c(NA, 
-40L), class = "data.frame")

Upvotes: 1

Views: 1665

Answers (2)

eipi10
eipi10

Reputation: 93761

You could add stackgroups=TRUE. Does that give you the effect you were looking for?

ggplot() + geom_dotplot(data = df, 
                      aes(x = Group, y = Response, fill = Recovered), 
                      binaxis = "y", stackdir = "center", alpha = 1,
                      stackgroups=TRUE) + coord_flip()

enter image description here

Upvotes: 2

lawyeR
lawyeR

Reputation: 7654

Does jitter help?

ggplot() + geom_dotplot(data = df, aes(x = Group, y = Response, fill = Recovered), 
                        binaxis = "y", stackdir = "center", alpha = 0.3, position = "jitter") + coord_flip()

enter image description here

Upvotes: 0

Related Questions