Correlation Bipartite Network: Improve mapping between edge weights and stroke widths

[d-callan]

ill attach an ex of the issue... but sometimes, particularly if there are few edges, some edges will have similar weights but very different stroke widths. like:

its possible we might want to use annotated display ranges, in combination w user inputs re thresholding of subsetting etc, to decide over what range to map edge weights to when finding stroke widths? worth mentioning in dataviz, to try to find a 'rule' that works for most people trying to most things most of the time..

[asizemore]

discussed in data viz meeting: Let's try always keeping the range 0-1 and get feedback!