Analysis on credit score
GitHub LinkWe are on the fast track to a cashless society in the world today. In a cashless era, credit cards are a great payment method. Banks do not, however, accept every credit card application. When choosing whether to approve a card application, they take into account a variety of factors, including the applicant's age, credit history, income, and ownership of the property. We can comprehend how each of these characteristics affects one in being authorised for getting a credit card using this dataset.
We are given with the credit score and their characteristics for 12.5k customers over 8 months, giving around 100k rows.Our dataset has many null values and outliers, and we must reformat certain categorical attributes for analysis. In this project we explore how can one maintain a good credit score, and what mistakes impact a credit score.
Understanding the data
Majority of Credit Scores are in the standard segment whereas there are less than 20% customers having a good credit score. The next step was to clean data - check for unexpected values in the column. For example credit history age which had time-periods mentioned as ‘5 year and 6 months’ that needed to be changed to numeric columns as 5.5.
Handled null values, removed outliers and observed the following trends:
Trends in Data
• Credit scores are almost equally distributed among all. So, it indicates that your credit score doesn’t depend significantly on your profession 
• Number of customers with good credit score tend to increase slightly month over month 
Factors affecting credit score
Effect of in-hand monthly salary
Effect of delaying payment from due date
• As the payment is delayed from the due date, outstanding debt increases. • People with good credit score and high outstanding debt clear the dues early.
• Credit History Age decreases with increase in payment delay
Effect of Outstanding Debt
• Customers in good credit score bucket, usually end up having lower outstanding debts and their monthly account balance is decently split between 0 to 1500 dollars • Customers in the poor credit score bucket has high outstanding debts and have low monthly income • Customers who have a longer credit history end up being good customers and they usually have less outstanding debts
Recommendations
• Maintain a high account balance and avoid delayed payments
• Maintain less than $ 1,500 outstanding debts to secure a good spot in good and standard credit score bucket
• Avoid having multiple bank accounts.
• Prioritize increasing the number of spends with large or medium dollar values and avoid payments with low small dollar values
