SASA Winning Team

Successful Turn at SASA 2019 Includes Analytics Simulation Challenge Win for UKZN Statisticians

At the 61st Conference of the South African Statistical Association, held at Nelson Mandela University, UKZN’s Discipline of Statistics attended in full force with close to its entire staff complement, while two young staff members and a postgraduate student triumphed as a team in the Cortex Analytics Simulation Game competition.

The annual event brought together local and international statisticians and researchers in related fields to share research, discuss new ideas, and meet and establish collaborations.

Dean and Head of the School of Mathematics, Statistics and Computer Science, Professor Delia North, spoke of the event as an important team building exercise for the Statistics Discipline, remarking that many attendees had noted the positive team spirit displayed by the UKZN delegation.

Masters candidate Ms Nina Grundlingh and lecturers Ms Danielle Jade Roberts and Ms Nombuso Zondo participated in the Cortex Analytics Simulation Game that ran for the duration of the conference, winning out over teams from a number of other South African universities.

The online game, a collaboration between SAS and HEC Montréal, presents a way for professors and lecturers to bring real-world experience into the classroom and enable players to learn analytics skills using predictive modelling technology. Developers also see it as having potential for participants to build their résumés through the experience they gain during play.

The game involves the posing of a scenario that is not unlike real-world problems. In this case, it involved work on a fundraising campaign for a charitable foundation, challenging competitors to efficiently raise as much money as possible by using predictive modelling software. Players received a data set of 1 million potential donors and used SAS Enterprise Miner to build a predictive model to identify the best donors to contact, uploading their results to an online leader board that ranked teams according to the most money raised.

Each team had only 20 hours available to work on the software programme and 25 attempts to attain the highest score. The team on the top of the leader board at the cut-off time of 9am on the final day of the conference was named the winner, receiving $1000 for their efforts.

‘It started off challenging as we were not all that familiar with the software as well as the predictive modelling techniques required to play the game,’ said Roberts. ‘However, with the support of one other on the team, we managed to find our way and come up with some great ideas to outsmart the other teams.’

The team was able to clinch the leading position on only their second attempt, from there on in having to beat their own score to stay ahead.

‘We learnt a lot in the process and can now say that we have developed some new skills in software and predictive modelling that will most certainly assist us in our careers,’ said Roberts. ‘In addition to these skills, we learnt a lot about how to work together as a team, playing to each other’s strengths. We believe these skills are vital in progressing our Statistics discipline at UKZN.’