I’d love to see topic discovery and link suggestion in org-roam; I’ve gotten used to something similar in Evernote, and I know I’ll miss it after I take the final leap.
Also, in case I’m not the only one who wasn’t aware of this, Hugo Cisneros posts his org-roam Notes in a graph displaying edge cardinality and topic clustering.
I think with what happened with GitHub - org-roam/org-roam-server: A Web Application to Visualize the Org-Roam Database, the introduction of GitHub - org-roam/org-roam-ui: A graphical frontend for exploring your org-roam Zettelkasten, org-roams built-in graphing abilities and generally users desire for better options in discovering and exploring their ZK, I would think that a standalone library that takes care of a bunch of useful algorithms like this could be justified. In then end the algorithms return org-nodes or their connections, but it would be worth communicating with the above projects what such algorithms can actually give them like the OP started doing, and coming up with a good API that can be used across the libraries. Otherwise people will start implementing a bunch of versions of the same base functionality like search and node/neighborhood discovery with slightly different APIs and possible with incorrect logic.
I have done some work analysing my network with the igraph package to analyse my Org-Roam database. Visualise Org-Roam Networks With igraph and R
Igraph has a large collection of algorithms with code in R, Mathematica, Python and C++.
Community detection (specifically the Walktrap or Spinglass methods) seem to work pretty good in knowledge networks.
@tecosaur and @danderzei
Thank you for this thread. Very interesting discussion. Do you have any updates about this?
I have two quick questions if you could help me with 