Stellenbosch University
Welcome to Stellenbosch University
SU excels in data science hackathon
Author: Daniel Bugan
Published: 15/10/2019

​The Transport Economics group within the Department of Logistics hosted a machine-learning hackathon at the LaunchLab in Stellenbosch on Saturday (12 October 2019), which saw a team consisting of Stellenbosch University (SU) electrical engineering students securing second place.

Although originally hacking refers to breaching network security to access protected information, nowadays hackathons are being held to find digital solutions to everyday problems using computer technology.

Uber and data science competition platform Zindi partnered with the department in hosting the hackathon which drew some 100 data scientists and aspiring data scientists from all over the Western Cape.

The objective of the hackathon was to build a machine learning model to help the South African National Roads Agency Limited (Sanral) predict when and where vehicle and pedestrian accidents are likely to occur in the Western Cape.

Participants were required to build their models on historic road incident data as well as traffic pattern data from Uber. The resulting models must enable Sanral to anticipate where they will be needed next and to put measures in place that will help ensure the safety of all Cape Town's residents.

According to a recent study by the World Health Organisation, approximately 1 million road accidents are reported in South Africa each year, resulting in approximately 31.9 fatalities per 100 000 people each year. This fatality rate could be dramatically reduced if authorities are able to respond to reported incidents as quickly and efficiently as possible.

The second-placed SU electrical engineering students – Matthew Baas, Ryan Seebregts and Kevin Eloff – said their solution code was based on a previously submitted code which they adapted and changed.

“This recognition gives us the confidence to pursue machine learning as a career. Machine learning is the future and this competition is a good opportunity to get into it while still studying for our degree," said Seebregts.

The trio received R2 000 in prize money for their efforts.

Marcus Gawronsky, a University of Cape Town student studying for his masters in Advanced Analytics took first place with R4 000. Third place went to Thembisile Damon, Xolisa Mzini, Siphesihle Yapi and Thabiso Mareletsa from Explore Data Science Academy who was awarded
R1 500.

Yolisa Kani, head of public policy for Uber South Africa, said government always complains that they do not have the resources to find answers to problems facing them.

“This hackathon shows that there are lots of people willing to use their talents to find solutions. I hope that we can extend this challenge beyond the Western Cape," she said.

Reginald Kgwedi, lecturer in the Department of Logistics, said Uber making their data available to SU “presents us with an opportunity to not only find solutions to the transport challenges in South Africa but to use utilise this information to find solutions in the academic space as we also need real-time information to use in our lectures".

Prof Stephan Krysgman of the Department of Logistics, said the hackathon is a wonderful opportunity for students to be exposed to practical examples, actual data and real case studies, and to work with an international company to solve a problem.

Celina Lee, CEO of Zindi, said the hackathon offers data scientists the opportunity to hone their skills on real-life data sets, solve real life problems, build up their professional profiles, and connect with potential job opportunities.

“The data scientists, especially the students, can use their experience in this challenge to get jobs in the field, they can even apply for a job at Sanral or even Uber," she said.

  • Photo: SU electrical engineering students (f.l.t.r.) Ryan Seebregts, Matthew Baas and Kevin Eloff secured second place in the hackathon.