My understanding is that the vertexnum will be an arbitrary integer
representing a user. For example, in the Twitter Small data set, the
first line reads:
45914942 25871621 5
Which means an edge is made from user 45914942 to user 25871621 with a
weight as 5. To balance things out, I used the vertexnums as keys for
a hashmap and then made the values an increasing integer, so you can
still organize your Adjacency Matrix in terms of simple integers while
maintaining the user vertex numbers.