During my internship at the Fundo Garantidor para Investimentos - FGI at BNDES, we faced the challenge of developing an internal risk model to compare the ratings provided by partner banks. When the beneficiary requested FGI's guarantee, the partner bank conducted a score analysis of that client and shared it with us. This allowed us to study the health of the portfolio.
Over a little more than a year, using both internal and external data, we created a logistic regression model in R that estimated the score of each potential beneficiary of FGI. Since it is an internal-use model, I cannot share many details here, but I am eager to talk about this delivery because it is very special to me.