R: ggplot2: position of ticks and labels in the absence of a grid

Question

This is a follow-up to this question: How can I suppress the vertical gridlines in a ggplot2 plot?

I have a histogram, like this:

library(ggplot2)
data <- c(1,1,3,2,2,2,2,1,4,1,4,4,1)
his <- qplot(data, geom="histogram")

I removed the vertical grid lines, using

his + theme(panel.grid.major.x = element_blank(),
            panel.grid.minor.x = element_blank()
      )

and now find that the ticks of the x-axis are too low, with a gap between them and the axis. While I can adjust the position of text using vjust=, there does not seem to be a way to adjust the position of the ticks, or does it?

Henrik's comment is of course correct. I just wanted to add that ggplot2 always adds a little space arround the data that you plot. This space can be avoided by adding the following to your plot `+ scale_y_continuous(expand = c(0, 0))`. — Stibu, Sep 18 '15 at 12:35

score 1 · Accepted Answer · answered Sep 19 '15 at 09:22

This is the example plot from the question:

library(ggplot2)
data <- c(1,1,3,2,2,2,2,1,4,1,4,4,1)
his <- qplot(data, geom="histogram") + 
         theme(panel.grid.major.x = element_blank(),
               panel.grid.minor.x = element_blank())

When ggplot calculates the range to be used for the axes, it alway adds a little space on both ends to ensure, that the data does not lie too close to the borders of the plot. You can prevent it from doing this by setting the expand argument of scale_y_continuous (and of other scales as well):

his <- his + scale_y_continuous(expand = c(0, 0))

This solves your problem of removing the additional space between the ticks and your data. However, you might want to have a little space on top of your data. To add this, you can use coord_cartesian as follows:

ymax <- max(table(data)) * 1.1
his + scale_y_continuous(expand = c(0, 0)) + 
      coord_cartesian(ylim = c(0, ymax))

I use max(table(data)) to get the maximal number of counts that the histogram will have and then add a 10% margin to that. Note that this way of fixing the y-axis range for a histogram works well in the simple case at hand, but in a more complicated situation, where the bins actually contain a range of values, you would need a more complex solution. Of course, you could always just create the plot and then read off an appropriate value for ymax.

This gives the following plot:

Let me also remark that the space between the ticks and the data has nothing to do with the precense or absence of the horizontal grid lines. So this solution works also if the grid lines are not omitted.

R: ggplot2: position of ticks and labels in the absence of a grid

1 Answers1