Grouping and filtering

Abstract Making sense of large and complex networks requires filtering and grouping as illustrated with the U.S. Senate co-voting network and CSCW Twitter network. Users can filter based on edge values, vertex metrics, and attribute data (e.g., Twitter followers). Even groups can be filtered out using the Visibility column. The Dynamic Filters feature allows you to interactively determine what is displayed on the graph by setting the position of sliders representing starting and ending ranges. Grouping vertices into subgroups (i.e., groups, clusters, communities, subgraphs) helps reveal important structures such as the Republican and Democratic divide. Subgroups can be created manually (e.g., using attribute data) or automatically (using one of NodeXL’s built in community detection algorithms). They can be visualized using unique vertex color and shape combinations, positioned using the group-in-a-box technique, or collapsed into single units or summary network motifs. Subgraph images can be created for each vertex.