Sean G
Sean G
Hi, how's going about this issue? Maybe I can provide some help if needed, because I'm familiar the d3-sankey API. 😆 BTW, I think sankey diagram designed for flow data,...
here is a demo: https://bl.ocks.org/mbostock/01ab2e85e8727d6529d20391c0fd9a16 but I discover it is still slow.
Solution: - #44 - #46
I think we need a new data structure. - https://en.wikipedia.org/wiki/Quadtree - https://en.wikipedia.org/wiki/Octree
![screen shot 2017-08-21 at 18 31 51](https://user-images.githubusercontent.com/3816105/29521787-8a9debb8-86b9-11e7-8f59-edac74547f51.png) ![screen shot 2017-08-21 at 18 36 20](https://user-images.githubusercontent.com/3816105/29521791-8f24ae4c-86b9-11e7-8315-3880a66cbce3.png) ![screen shot 2017-08-21 at 18 36 49](https://user-images.githubusercontent.com/3816105/29521800-92e9cd14-86b9-11e7-84ea-f43297ba6f00.png)
looks like ForceAtlas2 is faster than d3-force - https://github.com/d3/d3/issues/1936 - https://news.ycombinator.com/item?id=9864366 - http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0098679 - https://nblintao.github.io/ParaGraphL/
Too many points in one graph are of no value, my five cents.
yes, I've implemented it at https://github.com/netjson/netjsongraph.js/commit/1f3959ff33991f73f3282db1a5c9c1c889344bf0
@nemesisdesign now should add more tests and update docs
docs updated but examples