Data Quest 2024 Round Up

On Friday, Village Software held its fourth edition of Data Quest, a data mining competition for students.  Six teams were tasked with digging through large CSV files representing relative values and abundances of resources on an island.  The teams then race to plant their metaphorical flag on three small sections of that virtual island, to claim and sell those resources on a virtual market.

This year the Prosperity Award – the prize for the best performing team in the simulation - was claimed by Lilith Capps and Mengyun Li of team L.  Their use of a Long-Short-Term-Memory Recurrent Neural Network (LSTM RNN) helped them correctly weight the different resources and find some of the best spots on the grid.  Meanwhile, their use of parallel processing allowed them to secure those spots, being the first team to submit all three claims.

Bow Oyedepo, became single-person team Unprepared after being the one left over when forming teams from an odd number of competitors.  He handled this with both good grace and good strategy, acknowledging he likely wouldn’t have the capacity to contend for the grand Prosperity prize. Instead, he focused on trying more unusual approaches to the problem and brought forward some ideas that haven’t been seen in the competition before.  He was also quick to iterate, knowing when one method wasn’t working and switching to another. We felt that this got to the heart of the research and development process, and this, along with the fact that he still managed to come second in the simulation, garnered him the Innovation Award.

Byte Busters, Arman Irani and Mateus Padilha Luz, claimed this year’s Persuasion Award by virtue of the fact they remembered what the competition was all about – money.  While this seems like a trivial thing, they were the only team that saw past the numbers and code that had been filling the day and realise that all this analysis is done to drive people to make decisions. Other teams got caught up in the what of their solutions, but the simple fact of highlighting the financial implications of the exercise gave their presentations a grounding that clinical charts couldn’t deliver.

While these three teams claim the prizes this year, the day would not have been a success without the other competitors:

• Martin Adegbola Traore and Samson Nwizugbe of team Turing Machine, who also courted some more unusual approaches to the problem, but lost out in comparison with Unprepared
• Emma Phelps and Sarp Harbalioglu of team Smart Water, who suffered from Byte Busters’ tactic of rushing to get their first claim in as fast as possible – the first in fact to be submitted all day. Smart Water’s first claim attempt, submitted mere seconds later, had the misfortune of clashing with it and was hence rejected.  Had the submission order been flipped, Smart Water would have ended up in third place in the simulation rather than fifth.
• Arjun Krishna and Sowmmya Katari Hariharan of team Pheonix Force, who correctly identified the underlying mathematical nature of the problem in a way that most teams don't, but ran out of time before they could properly tune their algorithms.

Thanks to all of you.  We hope you found the day as exciting as we did.

We run this event because we love data, we love teasing out the meaning within it, and we love communicating those findings to others. We deploy business intelligence solutions to help our customers make the most of the data they produce and reach actionable insights so they can be more productive.  If you find yourself drowning in spreadsheets or stuck with disparate systems that limit what you know about your own business, why don’t you get in touch today.