search

pymc-devs / pymc

Bayesian Modeling and Probabilistic Programming in Python
https://docs.pymc.io/
Other
8.64k stars 1.99k forks source link

`forestplot` bug #551

Closed darribas closed 10 years ago

darribas commented 10 years ago

I think there is a bug somewhere in the forestplot method when quartiles=False.

I am following @fonnesbeck excellent example with the radon data. Say, for the case with varying intercepts, which is the first that uses the forestplot, I'm attaching the plots I get when I run the same command, only changing the parameter quartiles. That is:

forestplot(varying_intercept_samples, vars=['a'])

with

and:

forestplot(varying_intercept_samples, vars=['a'], quartiles=False)

without

It could very well be I'm missing something in how the plots are supposed to be (I'm still getting my head around them), but I would think the dots for the medians should be placed on the same locations under both options (quartiles either True or False). If not, a clarification would help me a lot to understand them. Thank you very much.

fonnesbeck commented 10 years ago

Thats a bug. Thanks for reporting.

fonnesbeck commented 10 years ago

Fixed in master. Please test.

darribas commented 10 years ago

Beautiful! I 'pip install git+https://github.com/pymc-devs/pymc' and it works as I'd expect now. Thanks!