In fact you just need to adjust a probit or logit model and then predict the score which are the "credit scores" you are looking for. If you don't have data, i recommend you to practice this:

https://archive.ics.uci.edu/ml/datasets/Statlog+%28German+Credit+Data%29

Greetings,

A very good research question. There was an interesting discussion. I believe that scientific research in this area should be developed. Greetings

In credit risk management processes, creditworthiness assessment procedures for potential borrowers and credit risk of banks, scoring methodologies based on many economic, market and financial indicators as well as index based analysis of quantitative data from surveyed entities are applied. In the portfolio management process, systemic credit risk in banks, the Value at risk methodology is commonly used. Credit risk modeling with R is a new approach that takes into account dynamically changing economic determinants of the economic processes under investigation.

Best wishes

Interesting for researchers working on modelling and big data.