Students from the University of Zululand’s (UNIZULU) Computer Science Department are in celebratory mode after being placed third in the Data Intensive Research Initiative of South Africa (DIRISA) Student Datathon Challenge.
The DIRISA student datathon is a competition that aims to showcase how open research data can be used to develop creative and innovative solutions to some of South Africa’s problems.
In this year’s challenge, student groups, who represent various participating universities, sought to outshine each other in finding a ground-breaking solution to South Africa’s Covid-19 issues. All students were expected to source open datasets and use these to meet the objectives of the theme.
The UNIZULU team members were Sinaye Sotashe, who is led the team; Mxolisi Khumalo; Lungelo Mpanza; Terrence Meluleki; along with their mentor and lecture Sizakele Mathaba.
Sotashe outlined that his group’s problem statement was centred around the South African Covid-19 vulnerability map. He said: “In many countries, census and other survey data may be incomplete and outdated. The 2011 census gives us valuable information for determining who might be most vulnerable to Covid-19 in South Africa. However, the data is nearly 10 years old and was updated 5 years ago, and we expect that some key indicators will have changed in that time. Can we infer important Covid-19 public health risk factors from outdated data? Building an up-to-date map showing where the most vulnerable individuals are located will be a key step in responding to the disease. We are also going to use data available on Stats SA to also train our model to make sure it performs well as desired.”
The findings of the group’s project exposed the population variabilities that exist in the country. According to Mpanza, the team was able to gather the that citizens who live in deprived conditions impact health and sanitation. Factors such as dwelling in crowded areas or informal settlements or even in multi-generational households impact social distancing, which ultimately increases the chances of the spread of the virus.
“We extracted South African census data and developed a state-of-the-art machine-learning algorithm that helps the government more accurately map Covid-19 risks in 2021, without requiring a new costly, risky, and time-consuming on-the-ground survey. Our main objective was to predict the percentage of households that fall into a particular vulnerability bracket using the South African census data. We used the algorithm results to calculate the total number of people who fall into the vulnerability bracket, excluding the already confirmed Covid-19 cases,” Meluleki explained.
For Mathaba, the commitment the students displayed while undertaking their project is a marvel. “I would like to thank them for the sleepless nights and for doing their outmost best in this competition. They have done the University proud,” she proudly shared.
- Precious Shamase