Use star symbols with ggnet2

[vertesy]

hi Shuangbin,

awesome package, thanks for building and sharing it!

Do you think I could use these symbols with ggnet2?
It has an internal shape = 19 parameter, but the final plot is a ggplot object.

Thanks
Abel

Example

library(GGally)
library(network)
library(sna)
library(ggplot2)

# random graph
net = rgraph(10, mode = "graph", tprob = 0.5)
net = network(net, directed = FALSE)

# vertex names
network.vertex.names(net) = letters[1:10]

ggnet2(net, node.size = 6, node.color = "black", edge.size = 1, edge.color = "grey", node.shape = 18)

[vertesy]

I used the description in ggnet/Hacking into internal values to achieve it by overplotting as on a regular ggobj.

[xiangpin]

The ggnet2 was implemented using ggplot2 layers and grammar. And the point layer was provided by geom_point. If you want to use the points of ggstar. It might need some non-standard process. You can replace the layer of geom_point of the final plot of ggnet2. You can refer to the following codes.

 library(ggplot2)                                                                                                                                                                                                   library(GGally)
 library(network)
 library(sna)

 library(ggstar)

 set.seed(123)
 net = rgraph(10, mode = "graph", tprob = 0.5)
 net = network(net, directed = FALSE)

 network.vertex.names(net) = letters[1:10]

 p <- ggnet2(
        net,
        node.size = 6,
        node.color = "black",
        edge.size = 1,
        edge.color = "grey",
        node.shape = 18
      )

# the node points is the second layer
 p$layers

# Since the character (size, color, alpha, fill) of node points was not used to map some variables in this example
# It is not need to use the mapping aes.
 new.star.point <- geom_star(fill="black", size=6, starshape=14)
 p$layers[[2]] <- new.star.point
 p  

If the character of node points has been used to map some variables, the mapping should be used now.

net %v% "phono" = ifelse(letters[1:10] %in% c("a", "e", "i"), "vowel", "consonant")
p2 <- ggnet2(net, color = "phono")
p2$data
p2$layers
new.star.point2 <- geom_star(mapping=aes(fill=color, starshape=color), size=6)
p2$layers[[2]] <- new.star.point2
p2 + scale_starshape_manual(values=c(13, 14)) + scale_fill_manual(values=c("#00AED7", "#FD9347")) 

But I don't think it is a good idea. ggraph might be a better option.

library(ggraph)
 library(tidygraph)
 graph <- as_tbl_graph(net)
 graph
 p3 <- ggraph(graph, layout="kk") +
           geom_edge_fan() +
           geom_star(mapping=aes(x=x, y=y, fill=phono, starshape=phono), size=6)
 p3 + scale_starshape_manual(values=c(13, 14)) + scale_fill_manual(values=c("#00AED7", "#FD9347"))  

[vertesy]

Fantastic, and very detailed, thank you!