Business Case: Detect fraudulent bank transactions
Finance Sector
A banking company processes thousands and thousands of transactions every week. Even with a very small fraud rate, a large loss can occur. Our customer has to fish out the conspicuous ones from the immense amounts of data - a case for our data engineers.
© Sascha Kohlmann: Withdraw Money, Licence CC BY 2.0
You shouldn't let anyone copy you at the ATM - businesslike bank fraud is even more subtle. With our data expertise, transactions can be checked for suspicious features.
Project requirements/tasks
Whether it's a movement of funds, the above-average use of a credit card or a direct debit from abroad - a bank must ensure that conspicuous credit card activities are checked through and criminal activities are excluded. Reliable fraud detection is an important security mechanism. To meet this requirement, our IT customers have to examine huge amounts of data: the entirety of all transactions with all their details. A task that can only be accomplished with modern data engineering methods.
This is our contribution
As part of a larger project team, we worked together to unravel the data packets and reveal potentially fraudulent patterns. In doing so, we brought our expertise to bear on this major project by using modern data processing tools such as the Hadoop cluster, the Spark engine and the Parquet storage format. Of course, we have always kept the important issue of data protection in mind: the DSGVO-compliant deletion of customer data is one of our services.
Our customer can now carry out a data-supported, functional fraud detection. He appreciates that with us he has a concentrated partner at his side who not only does the fair-weather work, but also finds his way around as part of a cooperation in a large team and provides the best possible support.
The project at a glance
Project title:
-
Fraud detection in the banking environment
Industry:
-
Finance
Project partners:
-
IT service provider
Term:
-
9 months
Our contribution:
- Data processing and data analysis
- Data deletion according to DSGVO guidelines
Technologies:
- Hadoop-Cluster
- Parquet
- Spark
- Java