Lending Club Case Study

The main cause of financial loss is giving loans to "risky" borrowers (called credit loss). When a borrower defaults on a loan or flees with the money owing, the lender suffers a financial loss known as a credit loss.

The major goal is to be able to identify these risky loan applicants so that the amount of credit loss can be decreased by reducing such loans. The objective of this case study is to identify such applications using EDA.

Analyze the variables that are reliable indicators of default in order to comprehend the driving forces (or driver variables) behind loan default. This information can be used by the business in each and every loan takers portfolio management and risk analysis.

Contact

Created by [@iamnarendrasingh] & [@chintan4560] - feel free to contact us!

Table of Contents

• General Info
• Data Processing
• Technologies Used
• Conclusions
• Acknowledgements

General Information

• The company aims to use exploratory data analysis to identify strong indicators of default, which can then be used for portfolio and risk assessment purposes.

Data Processing

• Data Preparation
• Correlation Matrix
• Data Analysis : Bivariate
• Data Analysis Univariate and Segmented-Univariate

Conclusions

• It can be observed that loans with a 36-month term are much more prevalent compared to loans with a 60-month term. Additionally, when looking at the loan status, it was found that customers who have defaulted have nearly an equal ratio of both term lengths. On the other hand, clients who have fully paid their loans are predominantly associated with the 36-month term. Thus, it is likely that borrowers with a 60-month term loan are at a higher risk of defaulting.
• The number of loan borrowers who own their home is very low, whereas those who are renting or have a mortgage are much higher in number.
• The verification status is a significant factor, as it is apparent that a large number of applicants are not verified. Furthermore, the proportion of charged-off applicants is similar in both the verified and not-verified categories. Therefore, modifications may be necessary in the verification process.

Technologies Used

• Python 3.X
• Google Colab
• numpy - 1.22.4
• pandas - 1.3.5
• missingno - 0.5.2
• seaborn - 0.11.2
• matplotlib - 3.5.3
• Windows 10
• VSCode

Acknowledgements

• https://stackoverflow.com/
• https://www.google.com/
• https://www.danielmsullivan.com/pages/tutorial_stata_to_python.html
• https://www.w3schools.com/python/numpy/default.asp
• https://www.w3schools.com/python/pandas/default.asp