Based on the Cross-Industry Standard Process of Data Mining (CRISP-DM), a loan data from Prosper is used to study key factors that predict loan Status. Specifically, I asked the following three questions:
- How do homeownership and employment status predict Loan amount?
- How do homeownership and employment status predict borrowers’ APR?
- How does Loan Status vary by homeownership status and employment status?
- A Descriptive Jupyter Notebook
- A README file
- NumPy
- Pandas
- Seaborn
- Matplotlib
No additional installations beyond the Anaconda distribution of Python and Jupyter notebooks.
Key results and findings were listed below. Find more on Medium
- For those with a home, I found that borrows' APR is the lowest for full-time employed.
- For those without a home, it is one of the highest for those who do not have a home.
- Putting together, it looks like those who are full-time employed and have a home enjoy the highest loan amount as well we the lowest borrower APR.
- Dataset is provided by Kaggle, an open-source data community.
- The analysis is benefited from the Udacity instructor and mentor team's help and support.