Posteriordb, version 0.3

Hi everyone,

Finally, there is now a new version up for posteriordb with some new content after some work by me @avehtari and @paul.buerkner and others. First, thanks to the work @kn2465, @andrewgelman, and @mike-lawrence, we have roughly 16 new, more complex posteriors thanks to @kn2465 s Google summer of code project. We still need to produce reference draws for these, which is the next step. Second, we have also gotten new models from @PhilClemson s group in Liverpool. Third, thanks to @stevebronder , we have improved the computational performance in some of the previous models.

There have also been quite a development on the R side to simplify the use of posteriordb from R, namely:

  • the R package has been moved to a separate repo stan-dev/posteriordb-r.
  • previous bug with running reference posteriors through R has been fixed (thanks @PhilClemson).
  • significant reductions in test suite computations (thanks @Bob_Carpenter).
  • many minor issues identified by @bbbales2 and @stevebronder has been fixed
  • new functionality to simplify contributions to posteriordb using R

Finally, we have also moved and set up the python library in a separate repo stan-dev/posteriordb-python, thanks to @ninpin.

Thanks, everyone!

The posteriordb can be found here.


Great. Will appear on blog tomorrow.