I remember this data-structure visualization system from a few years ago! Quick let me share some things I've learned since. Similar to how you're using color here, it's nice to have criteria for making nodes bigger than others. One gets a lot from strongly orienting things along axis. If you put this sort of visualization into VR, you'll be shocked by how many nodes you can fit in a space. And once that gets too dense, try implementing a sort of "breeze" to pass through the network, then the monkey brain will kick in to automatically disambiguate.