===================================================
The project aims at building logistic regression model that identifies risky loan (1 for good loan and 2 for bad loan) from bank details of its sampled 1000 customers.
The model produced the formula below to identify whether a loan is risky or not. If 1 then good loan, otherwise bad:
Here are the visuals of the variables from the dataset developed by creditData-insights.pbix file.
Here is a distribution of the amount of loan on whether it was classified as a bad (2) or good(1) loan by the bank.
The logit regression model produces a 2-shaped sigmoid curve below produced by the glm() function in R as seen in the credit-logistic.R file or credit_logisticAnalysis.Rmd file.
It is seen that most low amounts of loan (less than 3000) are considered good loan, while above 15000 is considered bad loan.