eulerr

To cite eulerr in publications, please use:

A BibTeX entry for LaTeX users is

Area-proportional diagrams with eulerr

This shiny app is based on an R package that I have developed called eulerr. It generates area-proportional euler diagrams using some rather groovy algorithms and optimization routines written in R and C++.

euler diagrams are generalized venn diagrams for which the requirement that all intersections be present is relaxed. They are constructed from a specification of set relationships but may sometimes fail to display these appropriately. For instance, try giving the app the specification A = 5, B = 3, C = 1, A&B = 2, AB&C = 2 to see what I mean.

When this happens, eulerr tries to give an indication of how badly the diagram fits the data through the metrics stress and diag error . The latter of these show the largest difference in percentage points between the specification of any one set combination and its resulting fit. It is the maximum value of region error , which is given for each combination. This metric has been adopted from a paper by Micallef and Peter Rodgers. Stress is more difficult to explain, but I advise the interested reader to read Leland Wilkinson's excellent paper for a proper brief.

I have listed some links on the right if you are keen on learning more about eulerr, perhaps even to contribute to its development. Finally, I owe a great deal to the aforementioned Wilkinson as well as Ben Frederickson whose work eulerr is built upon.

Johan Larsson

eulerr on the R package repository CRAN

My personal website

The Github repository for the R package

The Github repository for the shiny app