This is a very important reference: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4720993/
How is this different from UpSetR? Well a key thing is actually providing the data!
Figure 6: a) Exlcusive set (i.e. CCF + CKD but nothing else) vs. b) Aggregates which would include anyone with e.g. CCF+CKD
With exclusive sets we can determine the deviation which expresses 'surprise' of cardinality at a given intersection.
Cardinality = N Degree = N overlaps
Things to do:
- Clean up code to make more generalisable (?combineR) ~~ 2. Reduce redundency ~~ ~~ 3. Add tests for cardinality ~~
- Consistent documentation
- Implement deviation.
- ?implement upset visualisation