Chapter 6
Modeling with Bambi

A good tool improves the way you work. A great tool improves the way you think. – Jeff Duntemann

In Chapter 4, we described the basic ingredients of linear regression models and how to generalize them to better fit our needs. In this chapter, we are going to keep learning about linear models, but this time, we are going to work with Bambi [Capretto et al., 2022], a high-level Bayesian model-building interface written on top of PyMC. Bambi is designed to make it extremely easy to fit linear models, including hierarchical ones. We will see that Bambi’s domain is more comprehensive than just linear models.

We are going to learn about:

• Using Bambi to build and fit models

• Analyzing results with Bambi

• Polynomial regression and splines

• Distributional models

• Categorical predictors

• Interactions

• Variable selection with Kulprit