Zero-values for Directed Closeness Centrality

[emmalijohansson]

When doing the statistics for closeness centrality for directed networks (also betweenness centrality) I end up with a lot of 0-values. What is the reason for this?

I can imagine for nodes that only have 1 in-degree that they might get "trapped" when direction is taken in consideration. But I also get some 0-values for data with both in- and out-degree.

Is there an easy explanation?

[admin]

Hi,

Can you share your data in GEXF? It seems to be a bug.

[emmalijohansson]

Here is my network. It is weighted and directed, where the nodes are countries and the edges are the flow of land between countries.

[admin]

Hi,

I can't find 0-values for data with both in- and out-degree, but for the case of self-loop. Can you point me to an example? Thanks.

[emmalijohansson]

Examples are Congo and Costa Rica where Closeness Centrality, Betweenness Centrality and Clustering coefficient all are 0.

[admin]

Congo and Corsta Rica have out-degree of 0, so they can't reach any other node in the network. The results are therefore consistent.