Visa has a lot of focus on data science and almost all products with the organization aims to embed analytics in its core. Much of the work around data science within Visa happens from its India and Singapore offices. We decided to learn more from Amitava Dutta, who heads the Data Science function of Visa’s Consulting & Analytics team, supporting Australia, New Zealand and South Pacific market, and is based out of their Singapore office.
Amitava brings nearly 15 years of global experience in data science and consumer banking. Before Visa he led the data science team for payment products & consumer lending in Citibank India. Amitava resides in Singapore and holds a master’s degree in advanced Statistics & Data Analysis from Indian Statistical Institute, Kolkata.
Analytics India Magazine: How important is Data Science within VISA
Amitava Dutta: Data science, governed by strong data ethics and control, is at the core of everything we do at Visa. With over 110 billion payment transactions across more than 200 countries and territories, Visa’s business is built on huge amounts of data. The Visa Consulting and Analytics (VCA) team covers more than 20 countries and territories in Asia Pacific. The team consists of experienced data science professionals and business consultants. VCA helps clients with fact-based prescriptive solutions to drive their business goals.
Data science-led innovations are benefitting our clients from real-time fraud detection to designing value propositions suitable to consumers’ need. Merchants are leveraging Visa’s data science assets to provide uplifting shopping experiences for consumers.
AIM: Can you elaborate on a specific use case of data science that you worked on?
AD: Cross-border travel predict – At Visa, we have created an ensemble of machine learning models to identify which Visa cardholders might travel in the near future and what their likely destinations might be. Insights on travel propensity and likely destination(s) help our clients personalise offers. For example, we would recommend to our clients to identify customers with high propensity to travel with local offers along with other travel related solutions e.g. travel insurance, travel pre-paid cards, etc. After every campaign, performance is tracked against pseudo controls using our proprietary twinning module. Campaigns driven by such models have resulted in 28 percent increase in conversion rate for Travel Insurance cross-sell and 19 percent faster growth of spend at cross-border destinations for our clients in Asia Pacific.
AIM: How do you see the analytics ecosystem flourishing especially in south east-Asia region?
AD: Over 30 percent of the 650 million population in Southeast Asia are under 30 years of age and 90 percent of them have internet access. The emerging generation does not just know how to use technology but also how to build technology. Singapore is Southeast Asia’s leading innovation hub has moved up to the #3 spot on Bloomberg Innovation Index1 for the first time. Other notable Southeast Asian countries on the list include Malaysia (#26) and Thailand (#45). Against this backdrop, the demand for data science and analytics solutions has grown and the industry is experiencing a shortage of talent as a result.
In response, almost all major educational institutes in the region offer data science courses. There are many social communities to discuss and cultivate data science. Universities invite guest speakers from large corporations like Visa to share their views on data science.
At the same time, employees in the traditional business analytics field are encouraged to upskill themselves with machine learning courses, attend refresher courses on open source coding skills amongst other data science self-help learning courses.
Packaging rigorous statistical analysis into action-oriented visuals for decision makers is becoming increasingly critical to drive business.
AIM: Is AI talent an issue in Singapore? If yes, how can we resolve this?
AD: A successful data scientist is required to apply rational business sense along with machine learning skills to ensure that outcome is actionable. Strong data visualisation skills are an added advantage too. Visa’s campus outreach programs bring experienced speakers to share industry insights and interact with students on ‘Life at Visa’ as a data scientist. Such knowledge sharing sessions are excellent segues for our graduate hire program to grow and nurture the next generation of data scientists.
AIM: What is the biggest trend in data science/ AI that you look forward to in 2019?
AD: Data science is vital in today’s business and deeply entrenched in our business KPIs. The following are 5 trends worth watching out for:
- Data governance – Rapid adoption of digitisation and technology is going to create a larger data footprint than ever before. Data privacy and security are going to be at the forefront of data science conversations. At Visa, data science insights are governed by strong data ethics and control.
- Infrastructure – Right infrastructure, to solve Big Data problems. GPU computing is going to be increasingly popular to accelerate performance of machine learning models
- Data visualisation – More focus on Data visualisation to demystify data science and translate it into actionable insights to the C-suite executives
- Contextual marketing – Smart loyalty solutions powered by data science. Tailor made offers to engage customers catering to their personalised needs and aspirations
- Mastering the Black Box – Options for deep learning and complex neural networks with scores of parameters, hyper-parameters, and nuanced layers are usually quite dense. Escalating regulatory demands will focus on renewed efforts towards generating both the ability to explain and interpret black-box models