ArviZ is a Python package for exploratory analysis of Bayesian models, intending to provide backend-agnostic tools for diagnostics and visualizations of Bayesian inference in Python. Generally, Bayesian inference generates numerous datasets that represent different aspects of the model. To simplify the handling, referencing, and serialization of data generated during Bayesian analysis, ArviZ uses a couple of data structures: xarray.Dataset, arviz.InferenceData and netCDF.

InferenceData objects are central to ArviZ and most ArviZ functions take InferenceData as input. However, its functionality is still quite limited. The main aim of the project is to extend some methods from xarray.Dataset and to create specific InferenceData methods. Special attention would be given to test and document these new functionalities with examples.



Piyush Gautam


  • Ari Hartikainen
  • Oriol Abril Pla
  • Osvaldo Martin