refine graph creation about clics2 HOT 3 CLOSED

Ok, so far I considered every edge from a node in the graph to one not in the graph an out edge. But as you say, this could also be limited to edges with a certain number/kind of supporting colexifications.

from clics2.

For the archaic clics-visualization we used, it was useful to restrict it, as otherwise, every node has an out-edge and it gets less informative. If we decide to go for a cutoff of 1 word, this will definitely be needed. With three language families as threshold, it's actually a slightly different story.

from clics2.

In general, also with respect to #21, and the enhanced browsing of colexifications through the CLLD app, also allowing to account for missing data, it will be important to re-think the way the subgraphs are defined, specifically, how they are named. For the time being, they are named after the most central concept, but a subgraph is per definitionem built from one concept, from which we then go into all possible directions, so each concept has (if it is colexified) its own subgraph. At the moment, we have cases like:

• http://clics.clld.org/graphs/subgraph_312_END%20(OF%20SPACE)
• http://clics.clld.org/graphs/subgraph_346_END%20(OF%20SPACE)

where obviously the graphs are different but we don't know the original concept.

I think it will help if I write this up a bit more thoroughly in the paper draft. With @tmayer, we can later on then also discuss some additional ideas for visualization (coloring the central concept, maybe even adding articulation points [my favorite], etc.).

from clics2.