Geom_point and geom_errorbar not aligning to each-other

Question

I am new to R and have a problem with ggplot and the following dataset (chosen as representative from larger set) where geom_errorbar bars do not align with the mean point (using geom_point) and in several cases the horizontal bars do not align with the vertical bar in geom_errorbar, so that instead of coming out as an "I" with cross bars at top and bottom, the crossbars are separated from the vertical line or off-center.

I have looked at all the man pages for ggplot, geom_point, geom_errorbar, position_jitter (dodge, jitterdodge). I have also tried a bunch of things from here, such as altering the aesthetics within the geom_point and geom_errorbar calls (e.g. How to make dodge in geom_bar agree with dodge in geom_errorbar, geom_point)

Here's a basic data set:

df <- structure(list(
Test = c("A", "B", "C", "D", "A", "C", "D"), 
mean = c(1, 100.793684, 1, 1, 51.615601, 1, 2.456456), 
sd = c(1, 2.045985, 1, 1, 4.790053, 1, 4.250668), 
lower = c(2, 102.839669, 2, 2, 56.405654, 2, 6.707124), 
upper = c(0, 98.747699, 0, 0, 46.825548, 0, -1.79421)), 
row.names = c(NA, -7L), class = c("tbl_df", "tbl", "data.frame"))

Now the code I am using:

subplot <- ggplot(df, aes(x = Test, y = mean)) +
  geom_point(aes(x= Test, y = mean), 
             position = position_jitter(width = 0.2, height = 0.2))+
  geom_errorbar(aes(ymin = lower, ymax = upper),
                width = 0.1,
                position = position_jitter(width = 0.2, height = 0.2)) 
subplot

This is what I get:

I suspect it is something basic that I have missed. I have used the same code in line plots and other scatter plots and it has been fine, so I am lost as to what I have done. I have tested it on two different installations of R on separate computers too.

Any help greatly appreciated.

Answer 1

First,

Test = c("A", "B", "C", "D", "A", "C", "D")
mean = c(1, 100.793684, 1, 1, 51.615601, 1, 2.456456)
sd = c(1, 2.045985, 1, 1, 4.790053, 1, 4.250668)
lower = (mean+sd)
upper = (mean-sd)
range = 1:length(Test)

df <- data.frame(Test,mean,sd,lower,upper,range)

then

subplot <- ggplot(df, aes(x = Test, y = mean,group=range)) +
  geom_point(position = position_dodge(width = 0.2))+
  geom_errorbar(aes(ymin = lower, ymax = upper),
                width = 0.1, position = position_dodge(width = 0.2)) 
subplot

Answer 2

I posted this data set and problem to the ggplot Github page. It seems that I was indeed missing something simple - I needed to set seed for the geom_ calls to consistently jitter for each point. However it seems that there is an issue with geom_errorbar, as setting seed does not fix the crossbar problem.

Upon further investigation (from the Github team) it seems that the cross-bars are being jittered independently of the line. There is a work around (as of 23/10/18) to fix this. In the mean-time use position_dodge or geom_linerange.

  ggplot(df, aes(x = Test, y = mean)) +
  geom_point(aes(x= Test, y = mean), 
             position = position_jitter(width = 0.2, height = 0.2, seed = 123))+
  geom_linerange(aes(ymin = lower, ymax = upper),
             position = position_jitter(width = 0.2, height = 0.2, seed = 123))

Thanks to all for their help.

Answer 3

It looks like position_jitter is getting applied differently to the different components of the errorbars. That seems like a bug.

Here's a workaround that might accomplish your goals more directly. Add a column (I'm calling it version here) to distinguish between multiple runs of one Test, group by that column, and then use position_dodge to avoid overlaps.

library(dplyr)
df2 <- df %>% 
  group_by(Test) %>% 
  mutate(version = row_number()) %>% 
  ungroup()

subplot <- ggplot(df2, aes(x = Test, y = mean, group = version)) +
  geom_point(position = position_dodge(width = 0.5))+
  geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.2,
            position = position_dodge(width = 0.5)) 
subplot

Alternatively, we could use facet_grid and have the width change depending upon the number of tests, which will make the error bar widths consistent.

subplot <- ggplot(df2, aes(x = version, y = mean)) +
  geom_point(position = position_dodge(width = 0.5))+
  geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.2,
                position = position_dodge(width = 0.5)) +
  scale_x_continuous(breaks = NULL) +
  facet_grid(.~Test, space = "free_x", shrink = T, scales = "free_x")
subplot

Another approach would be to use a discrete scale, as you mention, perhaps by using interaction(Test, version) a variable combining Test and version to give the same width to each run. (I couldn't get the ordering to be by Test when using the interaction approach.)

df2 <- df %>% 
group_by(Test) %>% 
  mutate(version = row_number()) %>% 
  mutate(label = paste(Test, version)) %>%
  ungroup()

subplot <- ggplot(df2, aes(x = label, y = mean)) +
  geom_point(position = position_dodge(width = 0.5))+
  geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.2,
                position = position_dodge(width = 0.5))
subplot