igraph betweenness() is unnormalized (3.0) while networkx betweenness_centrality() is normalized (0.5) by default
On a 5-node path graph, nx.betweenness_centrality(P) returns {0: 0.0, 1: 0.5, 2: 0.667, 3: 0.5, 4: 0.0}, while igraph's Graph.Ring(5, circular=False).betweenness() returns [0.0, 3.0, 4.0, 3.0, 0.0]: networkx normalizes by default and igraph returns raw shortest-path counts, so tables, thresholds or rank blends that mix the two libraries are scaled wrongly. nx.betweenness_centrality(P, normalized=False) matches igraph exactly. Closeness differs too on disconnected graphs (networkx applies Wasserman-Faust scaling by default; igraph uses reachable nodes only). Fix: pass normalized=False in networkx (or divide igraph values by (n-1)(n-2)/2 for undirected graphs), and document which convention each metric uses before comparing. networkx 3.7, python-igraph 1.0.0, leidenalg 0.12.0, Python 3.12.