Usually a person is checked for credit worthiness using CIBIL Score in India and FICO in the US(It's a 3 digit number) which is based on repayment history, credit utilization, length of credit history, types of credit and many other quantitative data
But I'm checking if the person is credit/loan worthy based on alternative data sources such as rent, social media time, gig economy participation, online shopping and then using the given data itself to create a model to add a column called behaviour which could be based on psychometric analysis. The modified dataset is then used to predict whether the user is credit worthy or not
Model Accuracy - 99%