CODEWITHSUNDEEP
rdataframematrixigraph
pw225
pw225

Reputation: 51

Adding Attributes to Nodes in igraph based on matrix elements

I have created a network and would be interested in having the nodes possess attributes based on my initial dataframe. For example, my dataframe shows that each node as a certain amount of keywords but this is an information that I lost during the process of creating my network. At this stage I would be interested in having it attributed to each node for further analysis at a later stage.

Example of a graph:

g <- graph.formula(4506-8974, 8974-6345, 7842-4653, 4653-6345, 7842-8974)
V(g)$name <- c("4506", "8974", "6345", "7842", "4653")

The additional difficulty is that my dataframe takes the form of a matrix (although still technically a dataframe). The rows are nodes and the columns are keywords. I would like the columns (i.e. keywords) to become an attribute to each node: V(g)$keyword. If element = 1, than display as attribute, if element = 0, than don't take into account. Anyone knows how to do it? Should I use an ifelse function? Furthermore, I think it is important not to mix up the order of the nodes ID since otherwise it might give the keyword attribute to the wrong node.

df <- data.frame("agriculture" = c(0,1,0,0,0), "arts" = c(0,0,0,1,0), "banks" = c(1,0,1,0,0),
       "cities" = c(0,0,0,1,0), "companies" = c(0,0,0,0,1))
rownames(df) <- c("4506", "8974", "6345", "7842", "4653")

NodeID  agriculture  arts  banks  cities  companies
4506      0          0      1       0        0
8974      1          0      0       0        0
6345      0          0      1       0        0
7842      0          1      0       1        0
4653      0          0      0       0        1

 V(g)$keyword <- keyword based on if the element = 1 for each node

The result I'm hoping for would be for example:

V(g)$keyword[1]
[1] "banks"
V(g)$keyword[4]
[4] "arts" "cities"

Upvotes: 1

Views: 763

Answers (1)

Rui Barradas
Rui Barradas

Reputation: 76402

The following code does almost what you want. The difference is that V(g)$keyword becomes a "list". Extraction of its members will therefore be slightly different than your posted example.

V(g)$keyword <- apply(df, 1, function(x) names(df)[x == 1])

V(g)$keyword[1]
#[[1]]
#[1] "banks

V(g)$keyword[[1]]
#[1] "banks"


V(g)$keyword[4]
#[[1]]
#[1] "arts"   "cities"

V(g)$keyword[[4]]
#[1] "arts"   "cities"

Upvotes: 1

Related Questions