Skip to content

Analysis on the Prosper loan dataset using python and jupyter notebooks

Notifications You must be signed in to change notification settings

0xhaisenberg/prosper-loans-data-analysis

Repository files navigation

(Analysis of Prosper loan data)

by (Kosiso Aniebue)

Dataset

The dataset contains 113937 rows and 22 columns. Most of the columns are numeric but we also have categorical variables(both ordinal and nominal) and datetime variables. However, I will be working with just 84853 rows and 11 of those columns. The dataset can be found here, with feature documentation available here

I had to separate the LoanOriginationQuarter column so as to get the raw quarter values without the year attached

Summary of Findings

Prosper projects to averagely make 5-15% return on loans.

Loan terms are either year (12 month) or 3 years (36 months) or 5 years (60 months) with most of the loans having a 3-year term

Most loans are taken in Q4 followed by Q1. That indicates that most loan are borrowed between October and March. Although the number of loans originating in Q4 is higher, Q1 is bigger in terms of volume. Seasoality has no effects on the loan amounts or the estimated returns

The data suggests that on average it costs customers roughly 2.5% of borrowed amount to maintain a loan annually.

Most amount of loans are taken by earners in the 1-74,999 range and is generally around 5-10k

Most borrowers have monthly income around 5000 and seldom above 20000

Majority of borrowers have verifiable income. However, whether the borrower owns a home or has verified income does not determine the estimated returns the bank is to make on the loan

Loan amount increases across the three terms with better ratings

APR decreases with better ratings while Loan amount increases with better ratings Having a higher Prosper rating increases the amount of loan a borrower can access and consequently lesser APR

Key Insights for Presentation

Prosper projects to averagely make 5-15% return on loans.

Loan terms are either year (12 month) or 3 years (36 months) or 5 years (60 months) with most of the loans having a 3-year term. Longer loan terms bring in the highest returns. Larger loans also have lesser APRs where the estimated returns are roughly around 5-15% range while the lesser loan amounts have higher APRs that are largely within 0-25% range.

Most loans are taken in Q4 followed by Q1. That indicates that most loan are borrowed between October and March. Although the number of loans originating in Q4 is higher, Q1 is bigger in terms of volume. Seasoality has no effects on the loan amounts or the estimated returns

APR decreases with better ratings while Loan amount increases with better ratings Having a higher Prosper rating increases the amount of loan a borrower can access and consequently lesser APR.

It costs customers roughly 2.5% of borrowed amount on average to maintain a loan annually.

About

Analysis on the Prosper loan dataset using python and jupyter notebooks

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published