Lesson 4: Identifiying critical varibles and interactions

In Lesson 4 we learned how to take a large tabular data set with both continuous and categorical variables to predict an outcome. My question is how do we take this model that we developed, and evaluate previously not understood connections/correlations within the data. i.e. how to de find the essential variables and variable interactions to the model as well as the importance/weight of these influences.

In real life I would use this to identify the controllable and uncontrollable variables to help find efficencies in a manufacturing process.