Rotman School of Management, University of Toronto

Main Content

Processing. Please wait.
TD Data Hackathon – “The Rise of Data”, 2020

The 3rd annual TD Data Hackathon hosted by TD Bank, the Rotman School of Management's FinHub and TDMDAL brings together UofT students from all departments and programs with the opportunity to apply their business and data analytics skills to a real business problem identified by TD Bank. This year, the problem statement was Can you predict the net change in the firm's stock price on the day after a quarterly earnings call?

The output from the hackathon has had the potential to inform the investor relations and corporate communications team at TD on the appropriate content and tone of the bank's quarterly earnings call. This year’s topic was an interesting topic for researchers and an important question for analysts and investors.

From Friday, February 28 to Sunday, March 1, 2020, 22 teams comprising approx. 100 students from the Rotman MBA, MFin, MFRM, MMA, and BCom programs alongside students from the UofT Engineering, Computer Science, Economics, Math, and Statistics programs participated in the hackathon.

The second-placed team included three MFin students Tiffany Zhang, Di Shan, and Huanxing (Kevin) Zhou.

In this post, Tiffany, Di, and Kevin detail their participation in the hackathon as well as their ability to apply learnings from the MFin program directly onto the case.

More specifically, Tiffany, Di, and Kevin touch on how they were able to quickly apply the knowledge learned in the Innovations in Finance course in the MFin program to solve a real-world problem and compete at a high level in this hackathon event.

Choosing to participate in a hackathon as an MFin Student

While most MFin students participate in finance case competitions, Tiffany, Di, and Kevin decided to participate in the hackathon after taking a course called Innovation In Finance, -taught by John Hull. Coming all from finance backgrounds and working in finance-related fields, this course enabled them to gain more insight into innovations such as machine learning.

In this course, they learned a lot about python and machine learning, and for class presentations, Tiffany had the chance to present on natural language processing. This is what helped spark Tiffany’s interest in the hackathon and made her feel more familiar with the hackathon topic since it was finance related.

She put together her team, inviting Kevin along with Di. Di was a graduate from the MFin program in 2017, who came back to Rotman specifically to audit the Innovation in Finance course.

The Innovation in Finance course offered to MFin students is focused on the application of new technology to financial markets. With Professor Hull’s course, it focuses on business students that work in professional environments and that don’t necessarily have a heavy quantitative background.

“Professor Hull’s course is how I became interested in natural language processing. Without his course, I think that I probably wouldn’t have signed up for the hackathon. His course was a stepping stone for me to try new things out.” – Tiffany, MFin ‘20

The final team member was Rooney, a Master of information student. Together they were able to work through a number of different models, applying knowledge not only from their classes but their understanding of financial markets developed in the workplace.

They applied their diverse backgrounds together and went to work.

From start to finish

On the first night, the team was given instructions on the hackathon. They needed to predict the one day change in stock price for about 464 companies that are listed in the S&P 500. In terms of information, they were given the earnings call script transcript, closing price, and daily returns from all those companies and time frame 2008-2020. What they needed to do is build a model to predict and then be judged on the mean squared error and accuracy of their model.

“We needed to come up with a clear and concise story as executives are obviously busy people and they don’t have time to listen to lengthy presentations. We needed our presentation to focus on what we found and what business impact will it have to TD and why it is important.” – Tiffany, MFin ‘20

The team worked through a number of different models, utilizing their technical skills and industry knowledge to identify and build a model from a finance perspective using python. They started with a basic CAPM model and then removed the market factor.

The team discussed the different features they wanted to in the model. One example was looking at the language the CEO uses in public announcements. The team identified the words and phrases that were positive and those that were negative, and whether there was a relationship between performance. They used natural language processing to aid in predicting market returns. They followed those thoughts into future engineering and isolating factors to isolate returns related to the transcript.

As the day went on, they created a base model and would run through scenarios, and then adjusted the models throughout the day based on results that they’d find in the model.

They would focus on adding in more features, industry-specific betas, and sentiment changes based on trends. As the model became more sophisticated, they saw their results change. It involved lots of trial and error for the team.

“It’s a great teamwork experience and that’s what happens in the real workplace- you have someone who has lots of ideas but you need to balance that out with technical knowledge to make it happen. That’s what helped us.” – Tiffany, MFin ‘20

As for the solution, the team utilized learnings from the MFin program by using similar thinking and applying it to the case.

Earning second place

“I think what earned us second place is the way we approached the problem from a finance perspective. The way we approached it was very realistic in terms of an actual work environment of how they approach these problems. This is because we have this work experience and understand how it would go in a real working environment. This is the exact approach that we took and that really stood out when we presented to the judges.” – Di, MFin ’17

The team attributes Professor Hull’s course to their success in the hackathon, by saying that prior to taking the course they would not even think about participating in such a competition.

“We knew nothing about machine learning prior to this course. It’s like another world to us. With Professor Hull’s course, it’s the perfect bridge for finance students to work with individuals with engineering, computer science, and statistical backgrounds. While we are not experts, we are able able to work with those who are, I think that’s a huge advantage and it’s all thanks to Professor Hull’s course.” – Kevin, MFin ‘20

Final thoughts

There is an increasing shift in the fields of data analytics and innovation in the workforce. In her workplace, Di is seeing machine learning being applied finance industry. Her decision to come back to Rotman to audit Innovations in Finance has helped her have a better understanding of how the finance industry is changing.

“I think that this competition and course itself is really great in the sense that it’s kind of mimicking the real world and bringing together people that have the business knowledge to work with technical people. The process of collaborating is so crucial as we’re focusing on what they’re doing and communicating in their language while bringing back their results to business stakeholders and generating meaning from the numbers.”  – Di, MFin ’17

© Rotman School of ManagementAASCB