51±¾É«

Our glasses are running over...with data

When global beer-making giant Heineken wanted to find out why its products weren’t being consumed more in restaurants, the company assigned a team of data analysts and data scientists to find out why, and more importantly, what could be done about it.

Answering the first question was simple enough, according to the case study first published on 365datascience.com. When out for a meal at a fine restaurant, people who consume alcohol generally order wine because it pairs readily with a wide variety of foods. Heineken wanted to offer an alternative to the default practice of asking for the wine menu, and sought to find ways to showcase beer as equally versatile to meal-pairing as wine.

This is where the data scientists came in. They were responsible for identifying common molecule pairings in popular restaurant meals that produce the desired taste effect (i.e. salty, savory) and cross-referenced these molecular groups with the ideal beer molecules to create a tasty pairing.

Essentially, they broke every beer-food combination down into data sets then programmed a machine to identity which beers best suited which ingredients. From there, marketing analysts conducted market research and categorized which pairings taste subjects in New York City and Paris found the most delicious. The results of these tests were broken down yet again and data analysts cleaned and processed the data in a way that could be visualized by Heineken marketers and sales reps as they went about the task of convincing restauranteurs and customers to put down the wine glass and pick up a pint.

This example demonstrates the dynamic and multi-dimensional nature of the type of work data scientists do as they look for more efficient ways to identify and collect raw data through machine learning. But this process only works if they are working with data analysts to parse actionable insight from previously incoherent groups of numbers.

The success of complex data-focused projects requires the expertise of many people, all working together despite different job titles, perspectives, and skills -- with the common goal of producing better data intelligence.

According to a report from , the data scientist career path is “probably the hottest career choice you can make right now.” The median starting salary listed on the popular website is generous ($108,000 USD) with a high degree of job satisfaction reported by respondents of the Glass Door survey (4.2 out of 5), and with no shortage of available positions (~300,000 in North America alone). Career prospects look positive as more companies invest in data-literate workers who possess a knowledge of machine learning.

Meet Karoly Szalay (BBA/BMath '16), Manager, GBP Analytics at Scotiabank

As a manager of analytics at Scotiabank, Karoly Szalay (BBA/BMath ’16) is at the intersection of many data sets needing to be sorted out and investigated. But, instead of finding ways to drive beer sales, he’s working with different units at Scotiabank to find compelling insights about the bank and its services.

“I find this more interesting because I’m not confined to a single data set,” says Szalay. “Right now, my team is relatively small but focuses mainly on global payment products as well as operating accounts and payment accounts for our small business corporate commercial partners.”

Szalay was first introduced to financial analytics and data science at his first co-op position.

“I worked in quality assurance at QNX. It was a vague title, but a lot of that role involved resolving issues by creating new metrics or developing key performance indicators. That position really helped me understand that I liked working with data. My ideal position is having other business units come to me to help them find solutions.”

As a double degree student at 51±¾É« (BBA – Finance) and the University of Waterloo (BMath – Statistics), Szalay combined his love of math with the operational side of business, and now works in a numbers-focused role that has a direct impact on the financial health of the bank.

“It’s very exciting to have a direct impact on the bank’s financials,” says Szalay. “You can see your contributions in real time. For instance, I can work with other units to come up with a new pricing tier, and right away we can see our work reflected in current plans with the new pricing points being used and that’s really exciting.”

All the data coming to Szalay’s unit are raw and need to be cleaned regardless of the project. Some data sets aren’t clear about what a field truly represents, so Szalay’s team delves deeper to uncover the relationship those data have to a wider data set they have collected. From there, they determine which connections yield the best insights and build the best business intelligence for the bank.

For anyone considering a similar career path, Szalay has the following advice: “I would strongly recommend getting a good foundation in stats because it lends itself to critical thinking. With a knowledge of stats, you can see deeper in a way that informs your critical thinking in a more convincing way.”

“I would also recommend having a foundational understanding of one or two coding languages, just to be comfortable working with them. You don’t have to go around automating everything in Python, but knowing how to access a file or some basic commands would be very helpful as you start your career.”

As for Szalay, he’s not done expanding his knowledge either and has recently enrolled in a master of management analytics program to learn the skills he’ll need for the next step in his career.

“My ultimate goal is to manage a team of analysts and data scientists who are more focused on statistics. I also want to get more comfortable with machine learning to pull better insights in a shorter amount of time.”