I wanted a quick and foolproof way for visualizing univariate distributions using all three of histogram, KDE, ECDF on the same plot.
The package is here: Neven Sajko / UnivariateProbabilityDistributionVisualizationCalculation.jl · GitLab
I would in particular like feedback on:
-
the terminology in the README (I hope I didn’t say anything wrong or confusing)
-
Weights/multiplicities for input data: adding this concept to seemed like the obvious thing to do, but I haven’t encountered it before. Does this have a (some other) name? Are there any conceptual issues with this I should be aware of? See the README for an explanation.
-
My functions take inverse bandwidth (
1 / bandwidth
) for KDE, instead of the bandwidth. I figured this would be more convenient, but I’d like more opinions on this.
EDIT: to try this out, just clone the Git repo to a local directory and then do ] add /path/to/cloned
in the REPL.