Improved graphics for co-occurence networks

[jklevy]

Requested feature

It may be valuable to rewrite the production code for textplot_network. Currently one can change the vertex_size argument
to change node sizes.  There is an example below from a group in Dalian. In the their paper they create various co-occurrence networks of high-frequency words within the bioinformatics literature (K = 500, E = 200 etc),  where "the size of the node indicates the value of the node degree, which means the number of neighbor nodes connected to this node directly; only 50 nodes were from the top 200 edges among the top 500 high-frequency words"

While Quanteda can capture the node size change with the feature-occurrence matrix perhaps Quanteda's co-occurence networks could be made even more visually attractive by carefully displaying the nodes and arcs in the most aesthetically pleasing manner

There is some attached pseudocode in this paper dealing with the co-occurrence network of high-frequency words in the bioinformatics literature and its structural characteristics and evolution. I believe there are three appendices dealing with preliminary steps

https://www.mdpi.com/2076-3417/8/10/1994/html

Use case

This is a ubiquitous feature which can be used in a variety of contexts

Additional context

I think this feature could help to make Quanteda even more useful to a wide range of scholars and practitioners

[kbenoit]

Thanks! Good idea.