I am trying to use images (e.g. country flags) in a ggraph network. I was looking to use ggimage's geom_image
, however I reckon that the function needs to be adapted to work with ggraph where we do not specify the x and y coordinates.
library(tidygraph)
library(ggraph)
library(ggimage)
r <- create_notable('bull') %>%
mutate(class = sample(letters[1:3], n(), replace = TRUE),
image = "https://upload.wikimedia.org/wikipedia/en/7/7d/Lenna_%28test_image%29.png")
ggraph(r, 'stress') +
geom_node_point() +
geom_edge_link()
r %>%
as_tibble() %>%
ggplot(aes(x = runif(nrow(.)), y = runif(nrow(.)))) +
geom_image(aes(image = image))
# I would like this to work:
ggraph(r, 'stress') +
geom_node_image(aes(image = image)) +
geom_edge_link()
It's true that ggraph
does not require you to specify X and Y coordinates, but this is a convenience. The variable names are x
and y
. You can be explicit:
ggraph(r, 'stress') +
geom_node_point(aes(x = x, y = y))
And these variables are available to all other ggplot
-related functions, including ggimage::geom_image()
.
ggraph(r, 'stress') +
geom_edge_link() +
geom_image(aes(x = x, y = y, image = image))