Credit Card
Behavior Analysis

Credit Card Behavior Analysis
Overview of Credit Card Behavior Analysis

Summary

Company Wanted to have a system to analyze customer spending patterns on the credit card. They wanted to use this system to give better deals to the customer in the future. Moreover, in this world where fraudulent activities have become more accessible than ever, they wanted to develop a fraud alert system to secure their customer's data and funds.

Client Requirement

Keeping this in mind, our client approached HData Systems to build a robust technology to understand the customer's spending habits on the credit card to offer whopping deals in the future. This would increase customer loyalty towards their credit card company resulting in more spendings.

Client Requirement of Credit Card Behavior Analysis

The Business Challenge

  • The credit card statement only displays where the expense is made, for instance, Starbucks. But it doesn't show on which product the expense is incurred, such as coffee, wrap, bakery item.

  • Moreover, if we talk about Starbucks, it offers various coffees, so it isn't easy to detect which kind of coffee, size, etc. However, this issue got resolved with predictive AI technologies that helped our team get into specifications.

  • The user might use their credit card at several places, making it tricky for the alert system to detect fraud.

  • Another challenge is not catching and sending false alerts as it would trouble the customer and look bad on the company's end.

Solution

We have created a system that sends an alert based on the user spending pattern and location-based data. We used Python with the customer's transaction history as the dataset and ingested it into decision trees, Artificial Neural Networks, and Logistic Regression. As we feed more data to the system, we should be able to increase its overall accuracy.

For connection with a customer's credit card, we use an API called Plaid. The user connects their bank and credit card with us. We would categorize their spendings based on the tensor flow artificial learning model. Users can set custom limits on the category to manage their budgets.

End-Result

The solutions developed by HData Systems were of great help to the credit card company. They could efficiently examine their customer's spending habits and conclude what offers and deals would attract them the most.

After implementing the latest system, the company made reports by the end of 3 months where the consumer spending on their credit card spiked significantly, resulting in higher profits.

Moreover, the fraud alert system built by our team played a huge role in catching malicious activities and avoided detecting false fraud alerts. All thanks to our HData's expert team that managed to pull it off despite a few challenges.

Technology Stack

python
django
aws

CONTACT US

Get in touch with us

captcha