New Users' Credit Loan Risk Analysis

• 

  Make Your Dream Come True

Inspiration

Help the bankers detect who could be possible fraud or bad users.

What it does

Create a method of isolating out future risky users based on data from past late payments/overspending (per) credit history. Use this method to create a credit limit and awareness in a system for these new users or to isolate these users out of the application process completely.

How I built it

We tried to find out the data set from online resources and used Python to create a program that analyzed the information. We converted categorical factors into numbers and found the correlations between those factors with the overdue payment.

Challenges I ran into

Our data is not enough to be tested in a wider variety of data because, in current data set, they are not enough information to be used. Our R-squared value is not very high but the p-values are very good. We think that our factors may be matters but they need more data to be tested accurately.

Accomplishments that I'm proud of

We finally analyzed all the information from the table, cleaned data, created the chart, did some statistical tests and have some better insight into the problem. Moreover, we also created a website that lets the bankers or users type their information based on the data we have now to see whether the new users can be possibly a bad credit card or not. These assumptions are just based on the data we currently have. It does not mean that our tests are accurate but it gives us a better sense of the problem.

What I learned

• Communicated and worked with other CS students. Improved Python skills and logical thinking. Moreover, we got to know how the bank analysis works through researching information.

What's next for New Users' Credit Scores Risk Analysis

• Try to find more data sets and play around to see if we can rise the R-squared numbers.
• Hopefully, we get something interesting.

Built With

• javascript
• python