Prosper Loan Exploration

by Dolu Solana

Dataset

In this project we would conduct analysis loan data gotten from Prosper Loans, a San Francisco based peer-to-peer loaning app. The data consists of information such as borrower's rate, occupation, creditScore and other information.

Summary of Findings

We examined data after the 2009 when the prosper scores were created. We found several interesting insights such as:

1. Loan terms are usually one of three periods: 12 months, 36 months, and 60 months.

2. While Prosper Scores seem to have a relationship with traditional credit scores, it seems to be non-linear i.e A higher credit score doesn't always correspond with a high prosper score. A more indepth analysis will be needed to conclude on what factors directly impact prosper scores.

3. Prosper scores also tend to have a relationship with loan amount i.e Low prosper scores have rarely have high loan amounts. Though this relationship is not linear.

4. We also observe that high loan amounts are really given out in for short loan terms, we observed that 12-month loan terms maxed out around 25000.

5. While we can not make definitive conclusions on whether some occupations tend to borrow more, we observe that the highest number of individual loans in our dataset come from executives. Additionally we note that the highest median loan amounts where taken out by Pharmacists.

6. Borrowers' rate seems to range from 5% to 35%. With the most common interest rate around 35%, and the median interest rate around 18.95%.

Key Insights for Presentation

We would explore:

1. How loan terms are distributed across the dataset

2. The relationship between ProsperScores and Credit Scores

3. The relationship between ProsperScores and Loan Amounts

4. The relationship between Loans Amounts and Prosper Scores after considering Loan Terms

5. How different Occupations borrow money