Handbook of Graph Drawing and Visualization (2013)
cs.brown.edu
cs.brown.edu
Unfortunately the most ubiquitous graph type, hierarchical graphs, are an NP hard problem. The chapter outlines the Sugiyama framework for breaking down the problem, but each step is still difficult. Expect to sit down with the chapter over several sessions before you will be able grok how to draw a useful graph.
https://libgen.rs/search.php?req=Handbook+of+Graph+Drawing+a...
wget https://www.toptal.com/developers/hastebin/raw/norurilawu && pipx install pystitcher && pystitcher norurilawu output.pdfhttps://dmkolobov.github.io/tidy-trees-demo/
Good stuff on this page, and the references are an invaluable collection of prior art.
Page 807 (social networks)
Key to sociograms:
Girl, Boy
Colored Girl, Colored Boy
Housemother or instructor, Male instructor
We do one more step for UMAP and other dimensionality reducers in practice: connect the nearest neighbors to turn any data table into an explorable similarity graph. Users, devices, proteins, accounts, events, etc link together and cluster when similar, and this reveals fascinating structural relationships when there is a reason they vary.
Takes just a few lines once you have a feature table, and there's a recent GH issue to automate that step too: https://github.com/graphistry/pygraphistry/blob/master/demos...
Could you expand on this with an example. Or is there a reference or algo that I could look at. What do you mean by having extra info to generate coordinates. I am currently looking into graph drawing algorithms for a network viz problem and part of it involves placing nodes in certain places and I am interested to see if what you are saying is something I could use
If node or edge attribute information is useful for clustering, such as not just account IDs but say entity or event sizes & time, you may want to see sub-cluster separation based on that. A naive initial UMAP approach (I think the OP's intuition) is to ignore your graph structure and run UMAP (or t-SNE, PCA, k-means, ...) on your independent entities. No graph data involved (initially). Then you will get a scatterplot of entities where those with similar values across their features are near each other. The notebook I linked shows how to compute the UMAP on such a table. Then, instead of showing a scatter plot, you can then draw edges between neighboring entities and play with it: it's an entity similarity graph! That's the more interesting part of the notebook, and a big step up over how we see most UMAP workflows go.
However, it sounds like you already have a graph, which is more structure for clustering than out-of-the-box UMAP (PCA, ...) leverages. When you also have a graph of edges between those entities, such as causal events, and maybe even weights, you can also combine the similarity graph + yours, and run the graph clustering on the combined result. The clustering would then use both the node feature similarity and your existing graph knowledge. The UMAP similarity edges are likely more correlative & speculative than your physical graph's edges, so when running graph clustering, it often helps to assign different edge weights based on your confidence in each edge.
There are more tricks you can play here. A big one for property graphs is propagating node/edge features to nearby nodes (ex: "total connected bytes/events/$/etc.") so that the UMAP has surrounding graph information available. This starts to generalize to graph neural nets (b/c you're doing label propagation), and something we're actively looking into and are happy to chat w/ folks about: feel free to email or swing by our Slack :)
Funny enough, if someone is into modern versions of these graph techniques, we have been building out an end-to-end GPU visual graph intelligence platform used by all sorts of security/fraud/misinfo/genomics/supply chain teams, and have 2 data viz openings precisely on this: fullstack data viz + WebGL engine. Info over at https://www.graphistry.com/careers .
Once you load in say 100K-1B fraud events or cancer proteins and try to make the experience fully interactive, GPUs are letting us tackle all sorts of challenges that were unrealistic back then. With all graph neural net stuff happening in the last ~2 years, our assumptions of what tools can enable analysts to tease out are changing yet again. Exciting days!
There's also dagre-d3 which uses sugiyama. I have a different react project that uses that. I think I like cytoscapejs better, although the animated transitions from graph state to graph state are a bit clunkier.
I'm not aware of anything better for open source web that allows you to edit/animate, but I haven't tried gephi/graphology/sigma.js.
1. https://www.yworks.com/products/yfiles-for-html/pricing/#/co...