I am working on a project where I have been provided a data set of auto loans for a specific period of time with many loan parameters such as credit score, debt to income, loan to value, original term, interest rate, payment amount, vehicles year make and model, probability of default, unsecured DTI, etc.

I need to identify patterns in auto loan charge-offs (vehicle surrenders/repo) . Anything I want at all.

I’ve come up with some ideas: basic comparison of a number of parameters between charge-offs and non-charge-offs. Interest accrual on loans prior to charge off vs the calculated loss of the vehicle and how large the offset is. Number of months loan was working with collections prior to default and changes made to loan payments.

I am just looking for additional ideas on what I could show with this data. I’m happy to provide additional information if anyone is interested.