howto:analyze groups in nodexl

This tutorial demonstrates some techniques for using clustering tools for analyzing the results of clustering tools in NodeXL.

rawdataimage01.gif

Graph Type Directed
Vertices 15110
Unique Edges 13079
Edges With Duplicates 611
Total Edges 13690
Connected Components 2432
Single-Vertex Connected Components 3
Maximum Vertices in a Connected Component 9084
Maximum Edges in a Connected Component 10048
Maximum Geodesic Distance (Diameter) 17

Entire dataset is a bit of a mess. Let's look at it hour by hour. Use Autofill Columns > Edge Visibility to show for Hour = 1 so we see only hour 1 RTs.

This graph has
Vertices 3597

Unique Edges 2939
Edges With Duplicates 175
Total Edges 3114

Self-Loops 1

Connected Components 672

and looks like this:

rawdataimage02.gif

Is there an easy way to eliminate the "lesser connected components" from the picture? We can do a component census by doing a group by component and then constructing a pivot table on the Group Vertices sheet

PivotTableDialog01.jpg
NodeXL-groups-pulldown02.jpg

Aaron Clauset, M. E. J. Newman, Cristopher Moore. 2004. "Finding community structure in very large networks." Phys. Rev. E 70, 066111 (2004) [cond-mat.stat-mech] DOI: 10.1103/PhysRevE.70.066111

Ken Wakita and Toshiyuki Tsurumi. 2007. "Finding Community Structure in Mega-scale Social Networks." arXiv:cs/0702048v1 [cs.CY]

M. E. J. Newman and M. Girvan. 2004. "Finding and evaluating community structure in networks." Phys. Rev. E 69, 026113 (2004) DOI: 10.1103/PhysRevE.69.026113