Transforming industries through data

Data scientist Chandrasekar Venkatesh is at the forefront of pioneering work in anomaly detection across climate research, financial systems, transportation, and more. He played a pivotal role in creating new technology that improves road safety. We had a conversation with Chandrasekar about the ways data science can drive business growth and foster innovation.

Why did you become a data scientist? 

I enjoy solving complex problems through analysis and machine learning algorithms, as well as helping businesses to make informed decisions. As a data scientist, I provide companies with the tools to develop new products, services, and solutions that can revolutionize industries. Data science is at the forefront of technological innovation. 

What are some significant innovations that you helped to develop? 

I helped to design an alert system to avert accidents on cable-stayed bridges and mitigate financial losses from unnecessary bridge closures. These massive stainless steel structures are susceptible to ice formation. During severe winter weather conditions there’s a risk of large pieces of snow or ice breaking off from the stays and falling onto vehicles.

My first project was a landmark bridge in Toledo, Ohio – a stunning construction spanning 8,800 feet across the Maumee River. Our research team integrated weather sensors, collating information from diverse sources and employing data processing techniques. This allowed us to decipher the conditions that might potentially result in accidents causing injuries and fatalities.

This tool helped to prevent many accidents and save thousands of dollars in potential losses from unnecessary bridge closures. 

What insights have you gained from this project? How has it influenced your trajectory in the field of data science?

I engaged in the comprehensive analysis and processing of data from various sources, including weather sensors, local weather stations, and airports. This enabled me to discern the factors that might lead to road accidents.

Interestingly, the knowledge I garnered from this project has seamlessly translated into my current role in the financial services industry. The ability to detect anomalies, as I did with cable-stayed bridges, helps in addressing the persistent challenge of credit card fraud. Whether it’s combating cybercrime, countering credit card scams, or mitigating other financial threats, data scientists can play a pivotal role.

The ability to identify anomalies within expansive datasets is a vital asset in recognizing fraudulent behavior. For context, envision a local bank managing approximately one million transactions a day. In such a scenario, the skill of anomaly detection becomes paramount, revealing that around 10 transactions from this massive pool could potentially be fraudulent.

How can data science forecast financial threats or consumer behavior? Could you share some insights? 

We can predict a wide range of outcomes and events by analyzing patterns and relationships within data. This includes stock prices, market trends, and economic indicators. Consumer behavior is also predictable. 

For example, based on the data, I think e-commerce spending will remain strong this year. Inflation and eroding consumer savings could put downward pressure on spending in 2024. Some retail categories will also be hit, although events and entertainment could continue to see vibrant demand through the end of 2023. 

It is also noteworthy that fraudsters continue to exploit financial institutions even though several new tools and models are developed every year. Phishing, smishing, and catfishing are major contributors to card-based fraud. Fraudsters keep coming up with new ways to trick victims and take over their accounts. With excessive usage of VPN and the limitations and restrictions on what data can be stored by apps, it is increasingly challenging for financial institutions to detect and prevent fraud without giving genuine customers a lot of inconvenience.

What are the new trends in data science? 

One important trend is incorporating responsible AI practices to ensure that technology is developed in ways that respect fundamental human values and rights. We need to create AI systems that enhance human capabilities rather than replace them, and train machine learning without compromising individual personal data.

Major payment networks have also announced that they are working on newer AI-based models to tackle fraud but there hasn’t been much movement in the broader market over the last few years. Both Mastercard and Visa have disclosed that they are working on models to catch family or friendly fraud which is the biggest growing sector of ecommerce fraud.

This is when kids use a parent’s card to make purchases, or when a family member or a caretaker takes advantage of an elderly person. This is also similar to first-party fraud, where the customer is the fraudster. 

Many customers will falsely claim fraud on their account after a case of buyer remorse. This has become more commonplace in recent years with the growth of online gambling, streaming services, and speculation on crypto trading platforms. It is very difficult to predict which customers will turn into  fraudsters but that is the new challenge for all financial institutions.