[PaulHoule]

People still upvote hairball graphs every time. Fortunately there is a cure:

https://cambridge-intelligence.com/how-to-fix-hairballs/

[ng-henry]

The hairball was much worse before. I used a lot of techniques from this paper [1] to make it look decent and a bunch of other heuristics based on other papers to make it look informative.

[1] https://jgaa.info/accepted/2015/NocajOrtmannBrandes2015.19.2...

[mikk14]

I'd give it a shot to make node embeddings with Node2Vec [1] and then reduce them to 2D with UMAP [2]. I think it could help breaking apart the hairball, assuming you have a nice clustered structure.

[1] https://pytorch-geometric.readthedocs.io/en/latest/generated... [2] https://umap-learn.readthedocs.io/en/latest/

[tauchunfall]

Also edge bundling can help. See the papers by Benjamin Bach et al.

https://aviz.fr/~bbach/confluentgraphs/

[3abiton]

Is graph data processing considered visualization style? It is changing the data, how can this be considered "visualization"?

[ng-henry]

We aren't changing the data, just changing how the graph is displayed.

[omeid2]

The relationship of data is part of the data.

Your visualisation tool may require it in a specific format, but it still about properties of your data.