Closed finncatling closed 3 years ago
Consider randomly perturbing data on age rug to avoid issues with individual identification
The dataset is too large for rug plots - they are difficult to read even with a high transparency, and the PDFs of the plots end up containing lots of vectors so load very slowly. Therefore switching to histograms
Addressed in #81
Especially on continuous features, this would lessen overinterpretation on data-sparse areas of the plots