Change Vertex Node colour for specific edge igraph R

Question

I am working with an undirected graph. I am looking to plot the graph, and colour those nodes that satisfy some criteria.

I have a dataframe containing edges that are bad and good. Good edges I set the column color = darkgrey, bad edgeds color = red

My edgelist looks like:

      from     to    color
1    54770  54771 darkgrey
2    54770  54775 darkgrey
3    54770  54776 darkgrey
4    54770  54774 darkgrey
5    54771  54775 darkgrey
6    54771  54776 darkgrey
7    54771  54774      red
8    54775  54776 darkgrey
9    54775  54774 darkgrey
10   54776  54774 darkgrey
11  110780 110781      red

(to do this I run the following line on a column dob_diff):

df$color <- with(df, ifelse(dob_diff <= 168), "red", "darkgrey"))

I can plot the graph:

g = graph.data.frame(df, directed=FALSE)
plot(graph)

I can return a list of the index of those edges that are bad by doing:

a <- which(df$colour == "red") 
> a
 [1]   7  11

How can I take these edges and then colour those nodes which are connected to the bad edges?

This question here indicates you can access a node colour by its number

V(g)$color[6] <- "red" 

How can I tell igraph to take a list of edges?

Edit: For example:

I can return a list of those nodes that are incorrect:

> from <- cluster_data_updated_illegal_combined[c("from")][cluster_data_updated_illegal_combined$illegal_dob == 1,]
> to <- cluster_data_updated_illegal_combined[c("to")][cluster_data_updated_illegal_combined$illegal_dob == 1,]

> from
 [1] "54771"  "110780" "20285"  "83962"  "88825"  "19695"  "33343"  "89446"  "90175"  "98032"  "115768"
[12] "1982"   "71327"  "75017"  "75018"  "89315"  "89316"  "106277" "106956" "111140" "32841"  "72276" 
[23] "111062" "111074" "1904"   "12148"  "17325"  "17412"  "18238"  "18368"  "19681"  "28041"  "30755" 
[34] "31915"  "32825"  "43791"  "54806"  "54808"  "87075"  "94814"  "97877"  "101124" "102311" "107659"
[45] "108778" "108906" "112029" "120840" "127396" "33371"  "90150"  "111095" "122230"
> to
 [1] "54774"  "110781" "20693"  "83961"  "88565"  "20338"  "33340"  "91043"  "90171"  "98035"  "115771"
[12] "1976"   "71439"  "75019"  "75020"  "89929"  "89929"  "106279" "107489" "111139" "32993"  "72272" 
[23] "111060" "111071" "2697"   "12090"  "17323"  "17413"  "20300"  "18561"  "20025"  "28046"  "30744" 
[34] "31587"  "33113"  "44067"  "54799"  "54803"  "87073"  "94816"  "97872"  "101122" "103480" "107679"
[45] "109013" "109256" "112246" "120842" "127401" "32679"  "90259"  "111028" "122228"

I then try to colour the graph red for those nodes but I get an error:

> V(gg)[from]$color<-"red"
Error in unclass(x)[i] : invalid subscript type 'closure'
> V(gg)[to]$color<-"red"
Error in unclass(x)[i] : invalid subscript type 'closure'

Answer 1

What you got:

Your edgelist looks like this:

df <- read.table(header=T, text=" from     to    color
1    54770  54771 darkgrey
2    54770  54775 darkgrey
3    54770  54776 darkgrey
4    54770  54774 darkgrey
5    54771  54775 darkgrey
6    54771  54776 darkgrey
7    54771  54774      red
8    54775  54776 darkgrey
9    54775  54774 darkgrey
10   54776  54774 darkgrey
11  110780 110781      red")

This is the graph:

library(igraph)
g = graph.data.frame(df, directed=FALSE)
plot(g)

These are the indices of red ("bad") edges in your data frame:

which(df$color == "red") 
# [1]   7  11

Your questions:

How can I take these edges and then colour those nodes which are connected to the bad edges?

Here's one way to color the nodes, which are connected to the red edges:

vertices <- union(
  head_of(g, which(E(g)$color == "red")), 
  tail_of(g, which(E(g)$color == "red"))
)
plot(g, vertex.color = ifelse(V(g) %in% vertices, 6, 1))

See e.g. ?head_of. There might be other/better ways.

Is there another column I can create with a different name that igraph can use to auto colour the nodes?

You can assign a color attribute to vertices V(g)as well as to edges E(g):

V(g)$color <- (V(g) %in% vertices) + 1L
plot(g)