New faces, same electric energy, and unrelenting passion—Phitopolis Datathon 2K25 is back! This year’s event is particularly special as it is held for the first time in the new office, bringing an additional layer of enthusiasm and energy to the competition. The new environment added a different atmosphere, freer to move around for open collaboration with their respective teammates and friendly cross-team banter from time to time. There was a bit of anxiety on the days leading towards the event, but a lot of excitement when the event started, with people settling in their seats, finding where their teammates were. Days prior, participants were briefed with guide videos for preparation, registration steps to be filled up, and resources being prepared in line with their CWIQ Bench and Datastore accounts. Finally, the time has come, it’s time for the Datathon to start!
With every team locked in with their brainstorming sessions, datathon participants stretched the limits of data-driven solutions, solving challenging problems with innovative thinking and perseverance. This year’s roster is composed entirely of graduate program participants, whose fresh ideas and analytical skills make for a compelling mix. Joining them are first-time datathon participants Gabi and Bianca, excited to test themselves and demonstrate their abilities in this high-pressure competition.
As groups get into their rhythm, concepts start flowing, and the true challenge of strategy, pace, and creativity sets in. As teams settle in, ideas begin to flow, and the real test of strategy, speed, and ingenuity unfolds.
Driven by passion, resolve, and a continuous flow of pizza and drinks, groups of three convened, ready for the challenges that await them. The atmosphere in the room was charged with electricity as individuals were informed about the competition processes, data sets, and important measures were explained, establishing the premise for a high-stakes struggle of ability and tactics. With success at stake, the challenge was simple: create the most precise model and beat the rest.
The objective is to create a model that makes predictions of daily stock returns from the given datasets. But prediction is only half the story—your model will be evaluated on key performance metrics such as annual share, returns, max drawdown, up/down ratio, and dollars traded. All contestants begin with the same unprocessed data, but the way you preprocess it is what differentiates each team’s chances of taking the cake. So what’s the key to victory? Intelligent preprocessing and feature engineering. After data refining, split time—train your model, test its validity, and show your method outperforms the others.
Guided by members of the quantitative research team, every group plunged headlong into discovery, identifying patterns, interpreting trends, and fine-tuning their models with precision. Every move counted, whether it was finding the best variables to combine or through thorough preprocessing of said variables, teams worked against the clock to drive up their scores.
This year’s data presented a new challenge, distinct from past datathons, but with a well-crafted, thorough Notebook as their reference, participants were able to adjust quickly. Creativity and collaboration became the forces driving each breakthrough, with ideas freely flowing and strategies adapting in real time. As the competition intensified, each tweak and adjustment brought teams closer to the ultimate prize—a model that would reign supreme.
As the last submissions came in, hours have passed, the mood was a blend of relief, excitement, and quiet confidence. With hours of focused data analysis, model calibration, and final minute adjustments, the teams had done their best. All participants gained rich experience, new learnings, and enhanced problem-solving capabilities.
Phitopolis Datathon 2K25 was a triumph of collaboration, innovation, and data power. It was a stage where technology clashed with imagination, where novices walked hand in hand with veteran players, and where every task broke limits.
With another datathon concluded, we can be sure of this: this isn’t going to be the last. The stakes have been set higher, and we can’t wait to experience what the next generation of data warriors will offer. Until then, keep digging, keep constructing, and catch you on the flip side at the next battle!