previous.authors.graph: Create a graph from conversation data
In kasperwelbers/network-tools: Network analysis tools

Description Usage Arguments Value Examples

Create a graph from conversation data, in which ties are drawn between authors and the authors that communicated within a given distance (lookback) before them.

1 2	previous.authors.graph(conversation, author, order.nr, lookback = 1, direction = "directed.up", once.per.conversation = F)

`conversation.id`	A vector representing the conversation authors were engaged in (e.g. a thread, meeting)
`author`	A vector with unique author names/ids
`order.nr`	Optional. A vector representing the order in which authors communicated within a conversation.
`lookback`	The number of previous author with whom a tie with an author should be drawn
`direction`	A character string, indicating the direction of the ties. If 'undirected', ties between authors go both ways. If 'directed.up', ties are directed from author to previous authors (only if order is given). if 'directed.down', ties are directed from author to later authors.
`once.per.conversation`	Logical. If TRUE, max one tie between two unique authors will be counted per conversation, even if an author communicated within the given distance multiple times within the conversation.

A graph object in the igraph format

d = data.frame(conversation=c(1,1,1,1,1,1,2,2,2,2),
             author=c('Alice','Bob','Alice','Dave','Bob','Bob','Alice','Bob','Alice','Bob'),
              order.nr=c(1,2,3,4,5,6,1,2,3,4))
g = previous.authors.graph(d$conversation, d$author, d$order.nr, lookback=1) # how many times did author.X communicate directly after author.Y?
plot(g, edge.label=E(g)$weight, vertex.size=V(g)$n.documents*25)
g = previous.authors.graph(d$conversation, d$author, d$order.nr, lookback=1, once.per.conversation=TRUE) # in how many conversations did author.X communicate directly after author.Y?
plot(g, edge.label=E(g)$weight, vertex.size=V(g)$n.documents*25)
g = previous.authors.graph(d$conversation, d$author, d$order.nr, lookback=2) # how many times did author.X communicate within two messages after author.Y?
plot(g, edge.label=E(g)$weight, vertex.size=V(g)$n.documents*25)