Africa DSI Programme trains the next generation of researchers and industry leaders

The 2020 Africa Data Science Intensive (DSI) Programme was hosted by the South African Radio Astronomy Observatory (SARAO) in collaboration with AIMS South Africa, Grailabs and the University of Portsmouth as part of the UK-South Africa Newton Fund Development in Africa with Radio Astronomy (DARA) Big Data programme, between September 2020 and January 2021.

The Africa DSI Programme is a human capital development initiative to increase the data science research capacity and skills on the African continent. The programme uses real-world problems to give participants hands-on knowledge of the latest algorithms and techniques in data science and artificial intelligence, deep insights on industry trends, network building and practical team skills used in business to facilitate transitioning to a data science role in industry, academia or through entrepreneurship.

The Africa DSI programme commenced in September 2020 with 16 participants (50% of which were women) chosen from eight African countries and an initial pool of 1,430 applicants, showing the huge demand for data science and artificial intelligence training in Africa.

In 2020 the Africa DSI programme was adapted as an online part-time course due to the COVID-19 pandemic mixing lectures and real-world projects. Lecturers for the course came from all over the world, including large organisations such as Airbnb, Cambridge, DeepMind, Google, IBM and Netflix, and young African start-ups. The online course consisted of four modules over five months with the final module ending in January 2021.

Participants completed the following modules: Module 1: Regression; Module 2: Computer Vision; and Module 3: Natural Language Processing. For the fourth module, participants could choose their own topics, and on 26 January 2021, they submitted their final projects, which included a blog post, project report and code file. Their final presentations took place on 28 January 2021. Final project topics included, amongst others: disease image classification in humans and plants; crop yield predictions; geospatial natural language processing; restaurant and music recommendations; using voice recordings to detect early-stage Parkinson’s disease; and automated summarization of company income statements.

The top five presenters were selected to give presentations to a panel of judges, which took place on 29 January 2021. This final presentation winner was Nathanaël Rakotonirina from Madagascar, who gave a presentation titled ‘Video Analysis’. He presented work done on object detection in various real-life business environments.

Malcolm Wright from Botswana was the overall top participant for the 2020 Africa DSI programme, followed by Fanamby Randriamahenintsoa from Madagascar in second place and Martin Page from South Africa in third place. Lead organiser of the Africa DSI Programme and Head of Data Science at SARAO, Prof. Bruce Bassett, said: “Seven of the top ten participants were women, and there were three or four people within 1% of third place, so it was very competitive.”

Sohana Singh spoke on behalf of the participants and said: “We are certainly all honoured to have been selected for this opportunity and I can confidently say that we have all grown during this experience. Unlike the usual scenario where you have lectures and participants follow along, in the DSI, we were given real-world problems to work on in teams of four. We had to achieve a minimum viable product in a relatively short amount of time. I thoroughly enjoyed this way of learning as it was much more challenging and allowed us to push ourselves and learn things about ourselves that perhaps we didn’t realise before or think we were capable of doing”.

There have been various local and international lecturers on the course and four tutors, Dr Nadeem Oozeer (SARAO), Emile Lochner (Soldersmith) and two DSI 2018 alumni Evander Nyoni and Emmanuel Sekyi supported the participants.

“We are excited to host the 2020 Africa DSI programme as part of SARAO’s commitment to the UK-South Africa Newton Fund DARA Big Data programme”, says Dr Bonita de Swardt, Programme Manager: Strategic Partnerships for Human Capital Development. “This initiative builds on our capacity development interventions in offering data science training and skills development that can only enhance Africa’s capability in responding to regional and global challenges by exposing participants to real-world problems. We expect the need for these type of training initiatives to continue increasing into the future as can already be seen through the number of applications received from across the continent”, she further adds.