Prosper Loan Data Exploration

The final project in "Advanced Data analysis" NANO degree from Udacity

Overview

• The data contains information about 113,937 loan lists from Q4-2005 till Q1-2014 through 81 variables classified into 3 categories, and the data provided by Udacity.
• Using Data Wrangling process [Gathering, Assessing, Cleaning]
• Using Exploratory and Explanatory Data analysis
• Visualize data using [ Univariate plots, Bivariate Plots, Multivariate Plots]

Summary of Findings

• When Borrower APR increases the lender yield increase also.
• Borrower APR has a strongly -ve relationship with prosper score
• 36 months loan term has the highest frequency and wider Borrower APR range
• People prefer to take loans with 36 months term and it does not depend on the income range
• Peoples take loans for Debt consolidation more than any category
• Students take the lowest loan original amounts
• People with high monthly rates taking high loans

Key Insights for Presentation

For the presentation, focus on the relationship between those variables (Listing Category & Loan Original Amount & Term & Borrower APR & Employment Status & Lender Yield & ProsperScore) so I started to explore them firstly with Univariate Exploration after that Bivariate Exploration and finally with Multivariate Exploration to write a conclusion about the relationship between them.

• The relationships show that the loans with higher borrower APR and Lender yield tend to have a lower prosper sore
• People with full-time and part-time jobs had a lower borrower APR in all loan status
• In each income range when the loan original amount increased the medina of 60 months term increase more than 12 and 36 months term medians
• It is interesting because the median of 12 months term of the green loans category is higher than the other two terms
• The range of 12 months term in the cosmetic procedure is largest than the other two terms
• Personal loan and student use loan taking 36 months terms only