jlcx
commented
8 years ago Due to the large number of edges in this graph, a straightforward visualization of the entire graph with existing tools is probably not feasible. In addition to the ideas in #1, here are some ideas for visualizing parts of this graph specifically:
- combine certain categories or types thereof
- figure out a way to exclude certain less-important or finer-grained categories
- include only inter-category links, excluding the categorizations of the articles themselves
- alternatively, start at the bottom, letting users see other members of the different categories an article is in (these could be much smaller graphs)
- explore this graph further, and see what subgraphs have reasonable numbers of nodes and edges
There seem to be about 2 million categories, which could potentially be viewed in Causalaxies even though it's not the kind of dataset we're generally targeting. It might require changes to ngraph.offline.layout and/or the use of ngraph.native.