Imagine you are given a dataset of customer usage metrics. How would you identify which accounts are at risk of churning?
