We are going to explore a story of the economic crisis related to mortgage in 2008. A prevalent default behaviors lead economic crisis, and we seek the explore what factors lead the default. By exploring the data, visualizing data to see big picture, and building models,we can fully see a relationship of different characteristics and defaulting mortgage behaviors. Overall, after our analysis and modeling, we believe that mortgage issuer should adopt different kinds of information to different models to measure the credit risk of specific mortgages. Three modeling methods all agree that GDP, FICO Score, and Loan to Value Ratio will play important roles in people’s behaviors of whether paying or not paying back their mortgages. As issuers and financial analysts, the aggregate change of these variables should be closely examine. By applying different models, different variables' importance are re-evaluated, and we also realized that forecasting modeling is somehow subjective, which make me believe that using different models to forecast are the right thing to do.