I want to understand how predictive analytics is used for credit scoring. What types of customer data and machine learning models are used to assess credit risk? Can someone also explain how banks use these predictions in real-world lending decisions?