The category you are in: Economics The datasets you used from us: The current Lender's Club dataset from their website. Note that it is the same source as the one provided from the organizers and simply contain more data. Sources for any external datasets used: Lender's Club dataset Your team members: Doug Shi-Dong
Won't be able to be present for presentation.
Basic plotting of Lender's Club data to figure out some trends in the data.
As expected, the interest rate of a given loan is proportional to the risk associated with the borrower.
Then, using information about an borrower's application, use some machine learning tools to predict whether or not that person would default.
Interestingly, one has to take into account that borrower's who default do not leave the company with a complete loss since they usually make some payments before they stop. In fact, On average 54% of the principal amount is paid back before defaulting. When taking in account the successfully paid loans and the ones who have defaulted, an equivalent 6.45% profit is made on the loan on average, which shows that the accepted loan application have generated a profitable income.
Finally, using various common Machine Learning algorithms, a prediction is made by taking into account the features of the borrower such as home ownership, loan amount over the annual income, annual income, employment length and loan term requested.