Lender Bender
Lending Club is a company that helps borrowers and investors access credit. This project will use public data from LendingClub.com topredict the probability of a borrower paying back a loan in full to determinewhether they qualify for a loan. We will use machine learning to determine theprobability and find correlations of given parameters.
Tools:
Ski-learn, Flask, MongoDB, Pandas, Bootstrap, Tableau
CSV File Meta Data:
credit.policy - 1 if customer meet the creditunderwriting criteria, 0 if otherwise
purpose - purpose of the loan i.e., credit card, debtconsolidation, educational, major purchase, small business, and all other
int.rate - interest rate of the loan (a rate of 2% will be entered as 0.02).Riskier borrowers are assigned a higherrate
installment -borrower’s monthly installment for loan
log.annual.inc - borrower’s logarithm of income
dti - borrower’s debt to income ratio (amountof debt divided by annual income)
fico - borrower’s FICO score
days.with.cr.line - numberof days borrower had a credit line
revol.bal - borrower’s revolving balance (unpaidamount at end of credit card billing cycle)
revol.util - borrower’s revolving line utilization(amount of credit line used relative to total credit available)
inq.last.6mths - inquiries of borrower in the last 6months
delinq.2yrs – number of times the borrower was 30+days past due on payment in the last 2 years
pub.rec - borrower’s number of “bad” records (bankruptcy, tax liens orjudgments)
not.fully.paid - 1 if borrower did not fully paid, 0 if otherwise