[weishuhn]

Good feedback. Visualizations are definitely the most requested feature and something I'm working on. It's a hard problem to distill so many papers (5-150k) down into a usable graph. I've made big ones that look amazing but which are not that useful.

I like the idea of interactive filters to allow the person to explore it visually. I'm hoping to have something people can play with in the next few weeks. I hadn't really expected the site to take off the way it has and so it's not really feature complete yet :)

[anigbrowl]

It's outside of my technical wheelhouse, and most of what I know about graph theory is from trying to get a better handle on my domain problems, and as I get older I'm increasingly reluctant to get stalled on toolmaking.

But what you're discussing is a frequent problem with graph visualization - it's very easy to end up with 'hairball charts' that may be meaningful to the person who generated them but only because of familiarity with all the steps it took to get there, and the more inclusive the graph, the more time eaten by clustering algorithms, the more clusters produced, and the more of a cluster...well you get the idea.

As you're at an early stage of development, perhaps this technique, which trades away completeness for clarity and is relatively novel, might let you leapfrog some of those problems: https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8896846

[weishuhn]

This looks super interesting! I'll take a look. Thanks.

[austinjp]

I've no idea about Rust, but in Python land you might want to check out Graph Tool and similar tools. As you note, layout of large graphs is not trivial.