Loan Prediction on LendingClub Issued Loans Dataset

This project was developed for the Visual Analytics course of the Bachelor's degree in Mathematical Engineering in Data Science.

DESCRIPTION OF THE PROJECT AND EXPECTED BENEFITS

  • Develop a classification model to predict if a future customer will pay back the loan
  • Create a training dataset formed by internal data without outliers
  • Create a cluster method to identify main characteristics of customers that paid back the loan or not
  • Identify the most important causes that are related with customers that did not pay back the loan
  • Benefits: Assess whether or not a new customer is likely to pay back the loan.

REQUIRED DATA SOURCES

  • Customer data: Home ownership, annual income
  • Loan data: Loan amount, term, interest rate, grade, loan status, ...
  • Location
  • Propensity model
  • List of most important variables
  • List of characteristics to refuse/accept a loan issue

EXPECTED DELIVERY/OUTPUT

  • Dashboard in Tableau to discover the characteristics of customers that pais back and the ones that did not
  • Questions to answer: How are each customer type? Is there any relationship with amount/interest rate/term/grade?...
  • Map plots can be put together with bar, line, scatter plots