geom_tile in ggplot2 returning blank plot

Question

I am trying to plot a dataset with three columns using the code below, but I get a blank figure with just the names on the axis, can anyone tell me what I am doing wrong?

#rm(list=ls())
library(data.table)
library(wesanderson)
library(ggplot2)

exa <- fread("sample.csv", sep = "," ,header = T,  stringsAsFactors = FALSE)

pal <- wes_palette("Zissou1", 100, type = "continuous")
sp1<-ggplot(exa, aes(x=x, y=y, fill=z))+
geom_tile(show.legend = T)+
scale_fill_gradientn(colours = pal) + 
  scale_x_discrete(expand = c(0, 0)) +
  scale_y_discrete(expand = c(0, 0)) 
sp1

Data:

exa <- structure(list(x = c(0.457744, 0.492325, 0.49948, 0.471712, 0.504126, 
0.505528, 0.516318, 0.521419, 0.518354, 0.523001, 0.527301, 0.566585, 
0.411753, 0.427226, 0.446747, 0.43778, 0.432248, 0.444912, 0.462414, 
0.456952, 0.443462, 0.420683, 0.392057), y = c(25.69908, 25.65119, 
25.8332, 23.88982, 22.41502, 22.27553, 23.05898, 24.20714, 25.23666, 
25.35443, 25.78347, 27.20575, 20.94434, 21.62892, 22.61483, 22.49146, 
23.38523, 24.46414, 25.8023, 26.63754, 27.40164, 27.45981, 27.97814
), z = c(29.75408, 29.02752, 28.43744, 27.98952, 27.62504, 27.32658, 
27.04928, 26.77825, 26.53036, 26.28481, 26.04137, 25.80254, 28.5918, 
27.81994, 27.44201, 27.15059, 26.83333, 26.54576, 26.31531, 26.12999, 
25.88431, 25.65788, 25.47488)), class = "data.frame", row.names = c(NA, 
-23L))

ndichistan, I suggested an edit to your question that (a) comments out the previously-not-visible rm(list=ls()), it was not visible because your markdown code-fence was wrong, and please don't include code that "dangerous" so easily mis-run ..., and (b) removed the link to your data and replaced it with the data. If it's this small, it's much easier to just work with it here, and when links go stale (they almost always do), the question otherwise becomes unreproducible and less valuable to follow-on readers. Thanks! — r2evans
I cannot reproduce your problem: while they are small, I do see splotches sparsely throughout the canvas. Could it be that you're looking on a small plot window and just can't see the dots? Is your "real" data much larger, therefore things are being subdued for other reasons? — r2evans
Cannot reproduce either (although I didn't try your color pallete). — Axeman
@r2evans Thanks so much for the edits and suggestions for future posting. I did not see the splotches (maybe for the reasons you suggested), unfortunately, I cannot disclose the full data. It is however a 3cols*48rows dataframe. Is there a way to make it more visible? — ndichistan
Yeah there's just small tiles. Are you sure you want tiles? They need evenly spaced data, which your data doesn't look like. Can't you just make colored points instead? — Axeman

chemdork123 chemdork123 · Accepted Answer · 2020-06-11T20:51:57

Well, like others said, it's probably not the preferred way to represent your data in tiles, but rather use geom_point. In that case, here's something that works:

ggplot(exa, aes(x,y,color=z)) +
  geom_point(size=3) +
  scale_color_gradientn(colors = pal)

If you still want to use tiles, as others mentioned, you cannot see the tiles because theya are just too small. You can make them larger to see by specifying width= and height=. Since your scales for x and y axis are different, you can't specify the same number if you want square-ish tiles. For that, you can use a bit 'o math to do that for you. Here's something that works:

w= 0.02 * (max(exa$x)-min(exa$x))
h= 0.02 * (max(exa$y)-min(exa$y))

ggplot(exa, aes(x,y,fill=z)) +
  geom_tile(width=w, height=h) +
  scale_fill_gradientn(colors = pal)

BTW - the tiles there are still not square due to the aspect ratio of the plot.

geom_tile in ggplot2 returning blank plot

1 Answers