I have a data frame with more than 20 columns of actor names and I would like to make the adjacency matrix for the relationship among them, for the sake of simplicity let's say my data frame is as follows:

I would like an adjacency matrix in which I have an edge between Mary and Anna and bob and jack and also with Angela( basically, who ever is that is some how related to anyone)

I have already tried the following code but I will only get an edge between Mary and Anna(it put 0 for Mary and bob and bob and jack) and it seems like it only make the edge between first two columns.

ig <- graph_from_data_frame(data)
matrix <- get.adjacency (ig)

How can I get my desired adjacency matrix?

CodePudding user response：

I think you were missing an edge list; that graph_from_data_frame would need.

library(tidyverse)
library(igraph)
(start <- tibble::tribble(
   ~actor.1, ~actor.2, ~actor.3, ~actor.4,
     "Mary",   "Anna",    "bob",   "jack",
     "Lucy",   "Sara",    "bob",  "Alice",
  "Nicolas",    "rob",       NA,       NA,
   "Angela",   "Mary",   "Anna",       NA
  ) |> mutate(rel=row_number()))

(longlist <- pivot_longer(start,
                          cols=-rel) |> na.omit()|>
    mutate(value=factor(value)) |> select(-name))

(edgelist <- left_join(longlist,
                       longlist,
                       by=join_by(rel == rel),
                       multiple="all") |> filter(value.x!=value.y) |>
    relocate(rel,.after=everything()))

ig <- graph_from_data_frame(edgelist)
matrix <- get.adjacency (ig)