A correlation matrix is a square table created during bivariate analysis to show pairwise relationships and their strength across variables. Highly correlated variables can then be examined using business logic, with one potentially removed before dimensionality reduction. section 3.8
Use the matrix to investigate questions such as:
- Whether subscribers who use more data spend more time on the network than subscribers who send more SMS. section 3.8
- Whether online transactions generate more revenue than cash transactions. section 3.8
- Whether gender or age relates to data usage. section 3.8
