Table 1.
Properties of empirical networks and random networks1 with the same degree sequence.
Fig 1.
Assortativity rewiring algorithm with two pairs of nodes ((A,B) and (C,D) labelled according to degree and pairing).
A rewire will be performed to pair the highest degree nodes together, if not already paired, and therefore increase the network’s assortativity.
Fig 2.
Clustering Coefficient rewiring algorithm.
Node A is randomly selected and rewiring occurs if a path of length five exists and there is not already an edge connecting nodes A and D or C and E.
Fig 3.
Mean geodesic rewiring algorithm example where a pair of nodes with high geodesic is chosen and a rewire is implemented if it decreases the average geodesic in the network.
Fig 4.
Assortativity and clustering coefficient pair combination for the ANU friendship networks (N=217).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing assortativity through targeted rewiring increased the clustering coefficient.
Fig 5.
Assortativity and clustering coefficient pair combination for the FilmTrust networks (N=101).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing assortativity through targeted rewiring slightly decreased the clustering coefficient.
Fig 6.
Assortativity and geodesic mean pair combination for the EU institution email networks (N=610).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing assortativity through targeted rewiring increased the mean geodesic.
Fig 7.
Clustering coefficient and geodesic mean pair combination for the French primary school contacts networks (N=217).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing clustering coefficient through targeted rewiring had minimal or no effect on the mean geodesic.
Fig 8.
Clustering coefficient and geodesic mean pair combination for the US Congressional Twitter (N=475).
Increasing assortativity through targeted rewiring increased both the clustering coefficient and the mean geodesic.
Fig 9.
Mean (local) clustering coefficient and global clustering coefficient (transitivity) pair combination for the French primary school (N=153).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the clustering coefficient through targeted rewiring increased both the local and global clustering coefficient.
Fig 10.
Mean (local) clustering coefficient and global clustering coefficient (transitivity) pair combination for the Scottish corporate interlock (Boards) (N=131).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the assortativity through targeted rewiring increased both the local and global clustering coefficient.
Fig 11.
Mean (local) clustering coefficient and global clustering coefficient (transitivity) pair combination for the FilmTrust project (N=101).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the assortativity through targeted rewiring decreased the clustering coefficient and slightly increased the transitivity.
Fig 12.
Centrality measures under the assortativity targeted rewiring for the French primary school (N=153).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the assortativity through targeted rewiring slightly decreased the mean closeness centrality but not the mean betweenness or eigenvector centralities.
Fig 13.
Centrality measures under the clustering coefficient targeted rewiring for French primary school networks (N=153).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the clustering coefficient through targeted rewiring slightly decreased the mean closeness centrality but not the mean betweenness or eigenvector centralities.
Fig 14.
Centrality measures under the geodesic mean targeted rewiring for the Emails inside EU institution (N=610).
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Decreasing the mean geodesic through targeted rewiring slightly increased the mean closeness centrality but not the mean betweenness or eigenvector centralities.
Fig 15.
Mean and Gini closeness centrality pair combination for the ANU residence hall (N=217) network.
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the assortativity through targeted rewiring (left panel) slightly decreased the mean betweenness centrality and increased the Gini betweenness centrality. Increasing the clustering coefficient (right panel) decreased the mean betweenness centrality and increased the Gini betweenness centrality.
Fig 16.
Mean betweenness and Gini betweenness pair combination for the FilmTrust (N=101) network.
The gradient, lighter to darker, indicates the increase in number of rewiring attempts. Increasing the assortativity through target rewiring (left panel) slightly increased the mean betweenness centrality and decreased the Gini betweenness centrality. Note that the change in mean is extremely small. No relationship was found under the clustering coefficient target rewiring (right panel) between these properties.