Way to find longest string of connectivity?

Jan 12, 2011 at 9:31 PM

I'm creating a node that contains a list of patrons who attended a bunch of different events. Each of the patrons attended more than one of the listed events, and in many circumstances, many of them attended the same events.  The information has all been entered, but I'm wondering what the necessary steps are to 1) figure out the longest string of connectivity between the events via the people who attended them and then 2) isolate/showcase this data and only this data. Any help is greatly appreciated.



Jan 13, 2011 at 5:08 PM

I don't understand your scenario.  However, I can tell you that the only distance-related metrics that NodeXL can calculate are the graph's maximum and average geodesic distances, which involve the minimum distances between vertices (http://en.wikipedia.org/wiki/Shortest_path_problem), and those are just numbers calculated for the graph as a whole; they don't draw a path.  So I don't think NodeXL will do what you need to do.


Jan 18, 2011 at 2:43 PM

The Shortest Path Problem sounds like what I'm trying to configure. Another way to describe what I'm trying to achieve can be likened to the Kevin Bacon theory (http://en.wikipedia.org/wiki/Six_Degrees_of_Kevin_Bacon).  Imagine the graph to be a long list of actors and the movies they have been in, and you want to graph degrees of connectivity from Kevin Bacon.  I essentially want to do something similar with my list of patrons and the events they've attended, and want to figure out who my "Kevin Bacon" is and then determine the longest string of separation to/from him.  Hope this makes more sense.


If the Shortest Path Problem is what I'm trying to do, NodeXL won't be able to pull this off?


Thanks either way!

Jan 18, 2011 at 9:57 PM

No, I'm afraid NodeXL can't do that.  It will calculate the maximum and average geodesic distances for the graph as a whole, but it will not compute any distance metrics for individual vertices or pairs of vertices.

-- Tony