Heatmaps outside geom_ttile

[fwunschel]

Hi All,

Have loved all things geom_sf has done in advancing ggplot2, and have done some work with the geom_tile() function. I was wondering if there were plans to add a more fluid heat map (thinking a geom_smooth version for points) option or integrate sf objects in geom_tile(). If anyone has any tips or tricks to emulate a less pixelated heat map, sin generating infinitely smaller tile size, that would be great as well.

Thanks for all the work on this!

All the Best,
Frankie

[teunbrand]

Hi there, can you point to an example of the thing you'd like to do in ggplot2, but is currently impossible/cumbersome?
I have a hard time imagining what a fluid heat map, or geom_smooth for points would look like.

[fwunschel]

Something like this heatmap, rather than this more pixelized example. The difference im seeking is in the fluidity of color transition, the overlap on a map and the use of spatial typed variables - but maybe I just need to decipher a smaller pixel size and use an x y configuration rather than geometry I can achieve something similar?

[teunbrand]

For spatial data, it might be more convenient to look into specialised packages for handling raster data, like {terra}.
In terms of colour fluidity, there is geom_raster(interpolate = TRUE) that can 'blend' adjacent colours. There are a few more constraints on geom_raster() than on geom_tile() though.

library(ggplot2)

p <- ggplot(
  data.frame(
    x = c(1, 2, 1, 2),
    y = c(1, 1, 2, 2),
    g = LETTERS[1:4]
  ),
  aes(x, y, fill = g)
)

p + geom_raster()

p + geom_raster(interpolate = TRUE)

^{Created on 2024-10-03 with reprex v2.1.1}

[teunbrand]

I'm going to close this because it doesn't seem like there is something to do in this issue for ggplot2.
For fluid heatmaps you can interpolate data before you plot it.