As you may have guessed, multivariate EDA involves more than two variables. Here is where you will definitely get creative. A great deal of practice and effort is needed, to use the visualization and numerical techniques we've talked about so far in novel and interesting ways.
Common multivariate EDA techniques include:
- Coloring a scatter plot by categorical feature
- Using another categorical variable in the boxplots
- Conditional or lattice plots: dividing the analysis by different categories
- Parallel plots
- Heatmaps
- Principal component analysis and related plots
There are, of course, many others that we don't have the space to cover here. We will just provide a few examples of multivariate EDA to get you started.
For this section, we will go back to our credit card default dataset. Let's begin by importing the necessary...
