Related/alternative solution that might be more flexible (but probably also much more difficult to implement):
Separately: I disagree that the current all-in-one graph view is useful for seeing clusters in a large graph, unless you happen to have a graph that has clusters that are very weakly connected, which is probably unusual unless you are using the graph for two very separate purposes that don’t intersect much. In which case the clusters are already obvious in the data, and the graph view doesn’t add much. I’ve been thinking a bit about writing up a “how to make a graph view useful” type post, might have a go soon…