Operation Case Analysis

Goal

The Goal of this Analysis to get Customer Behaviour, while taking credit card loans and reasons of doing Charge off, and find the efficient ways to decrease them, and decrease the Credit Risk

Result

Results are presented in a report with recommendations and follow-up steps. Based on this report, we can decrease risk in Credit and clear and well-founded decisions can be made.

Project Overview

What is FICO Score?

FICO, originally Fair, Isaac and Company, is a data analytics company based in San Jose, California, focused on credit scoring services.

What is Payoff Loans?

Payoff Loans are the loans usually take to pay existing credit card Debt, and make monthly installement payment to pay off.

Project Analysis

Analyzing Which Score is Good For Predicting Bad Account.

After Analyzing Correlation and different Graphs, I have concluded that FICO Score has a Negative Correlation with Bad Account, which Means it is Inversely Proportional to BadFlag (Bad Account), While Score1 and Score2 having Positive Correlation with BadFlag, means they are Directly Proportional to the BadFlag, the Efficient Score is Score2 According to Correlation Aspect, but if Score2 Increasing, the Chances to be Bad Account is increased by 33%, but there can be more factors for Increasing Score2, but FICO Score is Decrease with the increase of Bad Account Probability, it is About 19%, So. I have Concluded that FICO Score is Being more Efficient than Any Other Score

This is the Correlation Heatmap, Which Shows the Correlation between variables and Correlation between BadFlag and different Scores has been Measured, can be checked in this Graph.

Credit Card Balance & Inquiries Last Month

CreditCardBalance is not very Efficient for Predicting Risk, Because the Correlation of CredictCardBalance is nearly 0, But InquiriesInLast6Months is More Efficient in Predicting Risk, as Compare to CreditCardBalance. Correlation of InquiriesInLast6Months is Positive, Means, the greater Number of Inquiries In Last 6 Months, the Higher Chances to be Bad Account

Estimating the Number of Monthly Charge Off (120+Days).

Before Explanation, A charge-off is a declaration by a creditor (usually a credit card account) that an amount of debt is unlikely to be collected. This occurs when a consumer becomes severely delinquent on a debt. Traditionally, creditors make this declaration at the point of six months, but in our case it 4 month or 120 days,

In our Dataset, we days range from where deliquiencies starts (0-14, 15-29, 30-59, 60-89 and 90-119) after that there is 90% Probability a consumer will do a charge off.

In this line Chart we Can see that Number of Charge off is Increasing Every Month in 2020.

Estimating the Roll Rates of Each Bucket

In the credit card industry, the roll rate is the percentage of cardholders who become increasingly delinquent on their account balances due. The roll rate is essentially the percentage of card users who "roll" from the 60-days late category to the 90-days late category, or from the 90-days late to the 120-days late category, and so on. in simple words, person who is late in pay the loans.

Forcasting Data for 2021 Q1 from the data of 2020

Predicting data for 2021 Q1, by using previous data of 2020, to find how much delinquencies in 2021 can occur

Predicted Delinquencies of Q1 2021

Recomendations

I have some data driven recommendations are as follows:

• we should develop an automated calling system call on an Hourly or Half-day basis, so we can remind our customers more consistently

• In the Data, people with less than an Annual Income of $70,000 are more likely to be charge-off, so while dealing with those customers, we should have some limit, so they do not charge off.

• If we created some strict obligatory factors, for charge-off customers It will also help us to retain customers

See this Project on My Github :

Get Prepared Report:

Download Now

Get in Touch

If you want to ask me something, or want to work with me feel free to message me I will reply you in a while