I am working with a Causal Loop Diagram (CLD), a systems thinking tool that shows how the variables that define a system are causally linked (based on consensus). If we represent a CLD as a network, its edges are directed; they have a value of either 1, to show a direct relationship, or -1, to show an inverse relationship.
Recently, some papers have applied netowrk analysis to get additional insight, so that the CLD becomes more than an insightful picture. But most of them just use Gephi to calculate centrality measures. It's been pointed out in a recent paper that this is wrong, because the calculations treat direct and inverse relationships as equivalent.
Is NetworkX the right tool to analyse a networrk with signed weights, and (1) calculate centrality measures (2) create partitions/cliques, accounting for signs?
I have tried analysing the network with Gephi; I understand that the results are not valid.