ggplot2 plotting more points than in the data frame, geom_point + facet_grid

Question

I have some data and am trying to make boxplots with an overlay of jittered points. My issue is with the points, so we'll stick with that.

Here's the data:

> dput(test)
structure(list(var1 = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 
4L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 
6L, 6L, 6L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 8L, 8L, 8L, 8L, 8L, 8L, 
8L, 9L, 9L, 9L, 9L, 9L, 9L, 9L), .Label = c("A", "B", "C", "D", 
"E", "F", "G", "H", "I"), class = "factor"), var2 = structure(c(1L, 
2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 
4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 
6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 
1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L), .Label = c("V1", 
"V2", "V3", "V4", "V5", "V6", "V7"), class = "factor"), response1 = c(5L, 
6L, 5L, 5L, 5L, 5L, 4L, 6L, 6L, 5L, 5L, 6L, 6L, 4L, 1L, 1L, NA, 
1L, NA, NA, 1L, 1L, 1L, NA, 1L, NA, NA, 1L, 5L, 5L, 4L, 5L, 3L, 
2L, 3L, 1L, 1L, NA, 1L, NA, NA, 1L, NA, NA, 2L, NA, 3L, 1L, NA, 
NA, NA, 4L, NA, 4L, 5L, NA, NA, NA, 1L, NA, 1L, 1L, NA), response2 = c(2L, 
2L, 2L, 2L, 2L, 2L, 4L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 5L, 5L, NA, 
5L, NA, NA, 5L, 5L, 5L, NA, 5L, NA, NA, 5L, 5L, 5L, 5L, 5L, 5L, 
5L, 5L, 5L, 5L, NA, 5L, NA, NA, 5L, NA, NA, 5L, NA, 5L, 5L, NA, 
NA, NA, 5L, NA, 5L, 5L, NA, NA, NA, 5L, NA, 5L, 5L, NA), response3 = c(4L, 
5L, 1L, 1L, 4L, 1L, 1L, 4L, 5L, 1L, 1L, 5L, NA, 1L, 4L, NA, NA, 
NA, 3L, 2L, NA, 4L, NA, NA, NA, 3L, NA, NA, 4L, NA, 1L, NA, 3L, 
NA, 2L, 4L, NA, NA, NA, NA, NA, NA, NA, 2L, 1L, 1L, NA, NA, 1L, 
NA, 3L, 1L, NA, NA, NA, 1L, NA, 3L, 1L, NA, NA, NA, 1L)), .Names = c("var1", 
"var2", "response1", "response2", "response3"), class = "data.frame", row.names = c(NA, 
-63L))

I used reshape2 to melt my data for faceting/simiplification of plot command:

library(reshape2)
test_melted <- melt(test, id.var = c("var1", "var2"), na.rm = T)

And here was the plot I created:

library(ggplot2)
p <- ggplot(test_melted, aes(x = var1, y = value)) + geom_point()
p <- p + facet_grid(~variable) + coord_flip()
p <- p + geom_jitter(position = position_jitter(width=0.2, height = 0.2))
p

That yields this:

Looked normal enough, but then I noticed it seemed like there were more points per facet/factor level than there should be. I narrowed down to just one level for var1

test_subset <- test_melted[test_melted$var1 == "E", ]

nrow(test_subset)
[1] 18

summary(test_subset)
      var1    var2        variable     value  
 E      :18   V1:3   response1:7   Min.   :1  
 A      : 0   V2:2   response2:7   1st Qu.:3  
 B      : 0   V3:3   response3:4   Median :5  
 C      : 0   V4:2                 Mean   :4  
 D      : 0   V5:3                 3rd Qu.:5  
 F      : 0   V6:2                 Max.   :5  
 (Other): 0   V7:3 

So, we should have 18 total points plotted (7 for response1, 7 for response2, and 4 for response3. Let's try it:

p <- ggplot(test_subset, aes(x = var1, y = value)) + geom_point()
p <- p + facet_grid(~variable) + coord_flip()
p <- p + geom_jitter(position = position_jitter(width=0.2, height = 0.2))
p

I count 11 points in the response1 facet, 8 in response2, and 8 in response3.

This has got to be something silly I'm missing. I've done plenty of faceting with dot plots and just never had this happen (or never noticed!).

Things I tried

• Removing coord_flip()
• test_subset <- droplevels(test_subset) in case empty factor levels were messing with something
• Playing with facet_grid(~variable) vs. facet_grid(.~variable) vs. facet_grid(variable~) vs. facet_grid(variable~.)

As one last note, I get different numbers of points depending on whether or not I facet. With faceting, I get 11 + 8 + 8 = 27, if I remove the facet_grid(~variable), I get 23.

Thanks for any suggestions!

Answer 1

The problem is not due to faceting, it is due to the fact that use two geoms in your plot. So geom_point will draw your points all in one place, and then geom_jitter will draw them once more, at random positions. That's why you can see one more point in each plot.

If you remove the call to geom_point everything is back to normal :

p <- ggplot(test_subset, aes(x = var1, y = value))
p <- p + facet_grid(~variable) + coord_flip()
p <- p + geom_jitter(position = position_jitter(width=0.2, height = 0.2))
p