You are a bank employee with access to a labeled dataset containing information on approved loan applications and their default outcomes. Your objective is to develop a model to predict default rates for future credit applicants. What steps should you take to accomplish this?