Roadmap

Wishlist for Bambi 0.5.0

• Implement new families such as Beta and T and add usage examples #310, #311, #312
• Add a .info() or .describe() method to Model that prints model information similarily to print methods in R for model objects (example). We don't have the exact same thing, but there's a rich output when you print a model now.
• Make getting started section executable #293
• Add a .predict() method to Model. We first need to update formulae so that any used transformation has memory about values used in the transformation (e.g. scale(x) remembers the original mean an std values of x) #105
• Save and load model. Not decided if they should be methods in Model or separated functions. #259

Wishlist for Bambi 0.6.0

• Splines. This first requires to update formulae #214
• Gaussian processes. This may require to update formulae

Wishlist for Bambi 0.7.0

• Splines. Partially available after #408 is merged.
• Multinomial models #318
• Gaussian processes #214

Long term

• Revisit and expand tests in general
• Decrease our dependency on statsmodels
• Consolidate ArviZ integration
  • Document new functionality
  • Support new functionality (including loo-related diagnostics)
• Add/improve examples (this is always and ongoing effort, but we can say this has been achieved compared to the previous stage)
• Revisit default priors see #230
• Work on porting code from books
  • Regression and other stories
  • Statistical Rethinking
• INLA support
• Allow "R-side" covariance structures and covariance priors in general (for varying effects too) #110
• Bambi fails when p > n #278
• Add example of posterior predictive sampling (and or check) #252
• Add example of prior predictive sampling (and or check) #251