0

I am trying to understand how Gephi/NodeXL measures Betweeness and Closeness Centrality or rather why there is a difference in measurement.

A twitter network with 1004 nodes and 2314 edges was sent to me. I measured centrality both in Gephi and NodeXL (directed network). Unfortunately I got two different results:

Gephi:

Betweenness | Closeness

A: 358.0 | 1.0

B: 0.0 | 0.0

C: 0.0 | 0.0

NodeXL:

Betweenness | Closeness

A: 295472,785 | 0,001

B: 91827,372 | 0,000

C: 92674,065 | 0,000

At first I thought the network caused the strange results so I tried to measure the statistics using other networks. It turns out that the network was not the problem. Can anyone explain to me how this is possible?

BalusC
  • 1,082,665
  • 372
  • 3,610
  • 3,555
curious_goat
  • 9
  • 3

1 Answers1

2

Gephi used a modified version of closeness centrality (CC) so that the higher the CC value the closer to the center of the graph. This modified version is an inverse CC and also takes into account the graph size, where as the standard CC is not dependent on graph size and is normalized between graphs of different sizes.

As for Betweenness Centrality, I get the same output from NodeXL and Gephi.

adm
  • 354
  • 5
  • 17