justinbois/distribution-explorer

Tool for exploring probability distributions.

distribution-explorer's Issues

Wrong Gamma functions

Hi, first of all, thank you for the great work; I've always been looking for something like this. Secondly, the Gamma function here and afterward should have an ( e^{-t} ) instead of ( e^t ).

CDF gets messed up in discrete uniform when the lower bound is changed

Most of the CDF vanishes when the lower bound is changed for the Discrete Uniform distribution.

Suggestion for Enhancing Learning Materials with Numeric Examples

Hi Dr., I can't tell you how useful the material is :) It's brief, concise, clear, and easy to understand. If we could have a simple numeric example at the end, right after the live charts, using some of the numbers from the charts along with an interpretation, it would be extremely helpful for reinforcing the content. FYI, I find some examples above the stories difficult to grasp at first glance since they require some domain knowledge (mostly biological); this can be distracting and divert attention from the main content because one must first understand what's going on within the example, if we could have more generalized examples it would be super fantastic.

fabs -> abs

fabs is deprecated in Stan, replace fabs references with abs.

Vertical lines in Uniform distribution not present in JS-generated data

Allow tabs to have various axes on log scale

It would be useful to allow the user to toggle between log scales and semilog scales in the plots.

numerical instability in beta dist?

Playing with the beta distribution, I noticed strange things like the following:

This should be symmetric about theta=0.5, clearly not.
Basically whenever (the parameter) beta is <1, things blow up at the RH boundary (theta=1).
Zooming in, we see that the grid point theta=0 is being rightly excluded:

while theta=1 is not,

which I suspect is blowing up in expressions of logprob, of form log(1-theta). Not yet sure where the fix should be though.