Stellenbosch University
Welcome to Stellenbosch University
EMS student and team secures first prize in influential data science hackathon
Author: Daniel Bugan
Published: 05/10/2023

​​​​In a remarkable display of data-driven ingenuity and problem-solving skills, three students from Stellenbosch University including one from the Faculty of Economic and Management Sciences (EMS) emerged as the winners of the prestigious Standard Bank Corporate and Investment Banking Hackathon.

The hackathon, held recently in collaboration with Mobalyz (SA Taxi), the SU School for Data Science and Computational Thinking and the Standard Bank Group, challenged participants to construct behavioural profiles for different taxi drivers using four months of taxi data.

The SU team of Christiaan Hildebrand (EMS), Wicus van der Linden and Daniel van Zyl made use of a dataset containing over three million individual taxi telemetry data observations spanning four months. This data included variables such as speed, g-force ratings and acceleration to name a few. They also were provided with the insurance claims data of the different taxis. The team had to use this information to construct behavioural profiles for these taxi drivers without having explicit information on the different drivers for each taxi.

Following a Markov-inspired approach and making use of external datasets such as Uber data and integrated weather data, the team came up with a model that predicts the risk label of a driver after each trip based on the data from that trip. The model also provides the drivers and taxi owners with instant feedback and useful information that will save them a lot of money. This real-time information not only empowers drivers and provides feedback to the taxi owners, but also contributes to safer roads.

Hildebrand, a Bachelor of Data Science (BDatSci) student studying in the focal area of Statistical Learning, said the team relished the opportunity to compete in the hackathon.

“We knew that we would be working with real-world data which is something we don’t often get the opportunity to do, and that excited us the most. We enjoy a challenge and we felt that this hackathon would broaden our skillset as data scientists.”

The trio walked away with R22 000 in prize money, but Hildebrand said they consider this only a bonus.

“The true prize in our minds was the connections and experience we gained. Before and after we presented our findings at the hackathon, we met and connected with a lot of people from Standard Bank and Mobalyz, as well as various data scientists and people very high up in the industry. These connections are far more beneficial to us than the prize money. We also gained a lot of crucial experience by working with real-world data to solve a real-world problem.”

Prof Paul Mostert, BDatSci programme leader, said the integrated skill set that the trio obtained through the different modules in their respective degrees, played a pivotal role in their victory. Hildebrand, along with van der Linden and van Zyl, both HonsBSc students, followed the same modules in Mathematical Statistics and Computer Science and were registered for all the Data Science modules up to their third year.

“It gives us great satisfaction to see students apply techniques they have mastered through their studies, and that the content we offer at Stellenbosch University is relevant and applicable for the industry.”

Hildebrand concurred with Mostert.

“We would not have been able to accomplish this achievement were it not for the knowledge and problem-solving skills we gained from our Data Science modules, not even close. BDatSci is truly one of the best, most useful degrees in the modern data world. I am truly grateful for the amazing knowledge I have gained so far in my three years of study.”

The BDatSci degree, introduced at SU in 2021, offers students the opportunity to gain knowledge of foundational modules in the core disciplines of statistics, computer science, mathematics and data science. They will also be exposed to the latest technologies and concepts in the field of data science.