What do car accidents and financial fraud have in common, according to a data scientist?

A few years ago a prominent data scientist, Chandrasekar Venkatesh, analyzed weather anomalies to prevent road accidents on cable-stayed bridges. He now helps banks and financial institutions to tackle fraud by identifying abnormal consumer behavior.  

As part of a research team at the University of Cincinnati, Venkatesh developed a tool that alerts drivers to adverse weather conditions at a landmark bridge in Toledo, Ohio that could lead to road accidents. This tool helped in saving thousands of dollars in public losses from unnecessary bridge closures. 

It turned out that Venkatesh’s experience was applicable to the financial services industry, where many banks and corporations struggle to prevent fraud. By detecting anomalies, just like with cable-stayed bridges, Venkatesh tackles rising cybercrime, credit card scams, and more. How can data science help to prevent financial threats? 

Diving into a large volume of data

In his research on traffic road incidents, Venkatesh had to analyze and process data from multiple sources, such as weather sensors, local weather stations, airports, and etc. to learn which conditions might lead to adverse driving conditions. This experience set him up for success in preventing debit and credit card fraud.

Detecting anomalies in large volumes of data is crucial to identify criminal activity. Out of around one million transactions in the average local bank each day, 10 could be fraudulent, said Venkatesh.

“In the last five to six years, crime has changed drastically because of the introduction of chip cards,” Venkatesh said. “Previously, cards only had  magnetic strips on the back, and it was very easy to fake them. Criminals used to install cameras on ATMs to get information.”

A lot of fraud was eliminated simply by moving to chip cards in 2015-2016. But criminals moved online and focused on e-commerce. “Card information can be stolen on a bad website or through a data breach,” Venkatesh said. 

For example, this summer, hotel group Marriott International confirmed a data breach exposing guests’ credit card information. The hackers used social engineering to trick an employee into giving them access to the computer. 

“Mastercard, Visa, Amex, PayPal — they all have algorithms and programs preventing fraud, but criminals are one step ahead,” Venkatesh said. “Fraudsters use bots to attempt thousands of transactions per minute by guessing card information. That’s why we use data science to stop them.”

Developing risk scores and models

Most fraud prevention and protection models in the banking industry are based on customer transaction patterns. All transactions, spending patterns, average checks, and a number of purchases per day are a part of the consumer digital profile.

To develop risk models, data scientists such as Venkatesh analyze information and share their recommendations with the bank’s software engineers. The risk has to be estimated within seconds. There’s very little time between when a consumer inserts a card, financial institutions come up with the risk scores and have to make a decision on approving the transaction. 

“If you use your card in a grocery store where you always go, the risk score is going to be low,” Venkatesh said. "But, say, your average check at the same store is $40-50, and one day you suddenly make a $500 purchase. Then the risk score will go up.”

Data scientists look at thousands of variables in transactions. Choosing the right variables is key to detect fraud. “It’s similar to analyzing weather patterns for my previous research,” Venkatesh said. “There are thousands of data points and there are different ways to identify the ones that are important.”

Although the ways to collect and analyze data are similar, risk management for banks and bridges are different. For example, during the holiday season, some banks enable more transactions. They accept some potential fraud-related financial loss in order to provide a smooth customer experience for their regular customers.

“Bridges can’t allow ‘more transactions’ in bad weather conditions to generate more profits, because this would cause accidents,” Venkatesh said. “But if a data scientist makes a mistake with fraud detection, an organization can lose millions of dollars in a day. So the stakes here are pretty high, too.”

Image sourced from Shutterstock

This post contains sponsored advertising content. This content is for informational purposes only and not intended to be investing advice.