BUSINESS TECHNOLOGY MEDIA COMPANY
Companies
Sectors
Software

Young coders tackle real world problems

Read time 3min 40sec
Nemisa’s Data Science Innovation Hack 2021 winner, Tinashe Ndemera.
Nemisa’s Data Science Innovation Hack 2021 winner, Tinashe Ndemera.

The National Electronic Media Institute of South Africa (NEMISA) recently held its second data science hackathon. Over 72 hours, data science enthusiasts from across the country were tasked with mining data and using it to build solutions to help government make more informed policy decisions.

Nemisa is a digital skills training entity of the Department of Communications and Digital

Technologies. The hackathon was held in partnership with provincial governments and took the format of an inter-provincial competition. Participants initially competed with each other provincially before each province’s winners competed nationally. The datathon’s top three solutions came from KZN (1st place), Gauteng (2nd) and the Western Cape (3rd).

Part of Ndemera’s research involved finding a correlation between the likelihood of a business succeeding and its owner's level of education.

KZN chemical engineering Master’s student Tinashe Ndemera developed the winning solution – a model meant to help government allocate resources to alleviate youth unemployment. 

“Unemployment is ubiquitous but the resources available to deal with it are finite, so I asked myself how best those resources can be allocated to reduce unemployment,” says Ndemera. "The model will use factors such as the country’s education and entrepreneurship levels to calculate the likelihood of citizens finding jobs and then determine where resources can be routed to create employment opportunities."

Entrepreneurs create jobs, so part of Ndemera’s research involved finding a correlation between the likelihood of a business succeeding and its owner's level of education.

“I couldn’t find this data. What eventually helped was someone’s MBA research into the matter, with only 13 participants”. The data showed him that even some level of formal education helped sustain businesses.

“I also used information from a KZN local government skills audit. This clearly showed the lack of skills facing the provinces wards, so even if there were job opportunities, the youth are in no position to take advantage of these opportunities."

Ndemera‘s model won him the R15 000 provincial grand prize and R100 000 for product development for his national win.

Let data lead

This was Ndemera’s first hackathon. He came across the competition while doing research into machine learning and deep learning.

“I wasn’t sure what data science entails but I still entered. Through research I realised I’ve actually always been doing it through my studies. I’ve also only recently started coding in Python. The hackathon gave me a chance to apply what I’ve learnt.” 

He adds that his biggest takeaway from the hack has been ‘to let the data tell its story’. “Our mentors at Innovate Durban emphasised the importance of using data to build your solution. I was getting ahead of myself at some point, making assumptions on what I’d find.”

Lemogang Matlou is a 16-year-old coder from the North West, and this was also his first hackathon. Matlou is in grade 11 at JM Ntsime High School and while he’s studying maths and science, his school doesn’t offer IT. As an aspiring data scientist (or software engineer) he’s now learning about data science and coding in his free time. 

Matlou says a Facebook friend inspired him to get into data science. “I used to see Kutlwano Tshatiwa post a lot about it. Now we’re offline friends and we plan on meeting at Eskom’s Young Scientist expo later this year.” 

In 2019, Tshatiwa’s anonymous Web browser won a ‘highly commended project’ award at the Expo.

Matlou won first place in the datathon’s regionals for his Web app, SA Statistics-Reporting and Forecasts, that uses machine learning to make predictions on, for example, incidents of crime and unemployment. The app can be used by government and individuals.

“At first I didn't know what to do with the dataset I had, but while analysing it I saw that I can put a report together and then make predictions," he says. "Sometimes government struggles to make informed decisions; it can use these predictions to prepare for the future.”

See also