This past Saturday (12th August 2023), Predictive Insights hosted a hackathon at the University of Cape Town. The event was a resounding success, bringing together bright minds to tackle a pressing challenge in our country, namely unemployment.
In a setting characterised by limited resources, how can we allocate our limited resources strategically and gauge the true impact of interventions? Our Zindi challenge aims to find machine learning solutions that help answer these questions, and the UCT Hackathon was a pivotal part of this competition.
A standout moment of the hackathon was the insightful session led by Cobus Burger, where he shared some invaluable tips and tricks that set participants on the path to success.
Cobus reminded the students of the importance to truly understand the data – to delve into each variable, draw meaningful graphs, and perform initial correlation or regression analyses. He also emphasised the significance of constructing a robust cross-validation method, a cornerstone for accurate model evaluation.
Participants were encouraged to keep a keen eye on the “scoreboard,” actively engage in discussions, and observe the strategies of fellow competitors. The spotlight then shifted to practical ways of enhancing model performance, such as the use of feature engineering tools and ensemble methods.
Our sincere congratulations to all the winners! And a heartfelt thank you to everyone who participated, supported, and contributed to making this hackathon a remarkable success.
The Predictive Insights Hackathon was not just a competition, but a journey of exploration, learning, and camaraderie. The event exemplified the vibrant spirit of the data science community and left everyone excited for the endless possibilities that lie ahead!