Closeness Centrality

Aug 28, 2014 at 10:03 PM

I am a reasonably new Node XL user and have generated some metrics for the network I am looking at. The Closeness Centrality measure comes in as 0.014 for the central user (ego) and 0.007 for every other vertex in the network (73 other vertices). Now I realise that this figure is below 1 because the network doesn't show interaction between all vertices in the network, it only shows interaction between the central vertex and its surrounding vertices. But why is the measure higher for the central vertex and lower for the surrounding vertices?

I am looking at a 'mentions' network for Twitter, the central vertex mentions other vertices more than once (up to 19 times for one specific one), is this the reason why the number is higher?

I was a bit surprised because when reading the Node XL book on page 72 it states that 'The lowest possible Closeness Centrality score is equal to one.' Well these figures are below one?

I'm probably missing something really obvious but could do with a little guidance.