Big Data Mauritius
DARA Big Data has had two postgraduate astronomy students in its cohort from the beautiful island of Mauritius, David Bastien and Zafiirah Hosenie. Both of them were heavily involved in the organisation of the successful Big Data Mauritius event, held in late June 2022. The event also had the strong support of our partner organisations OAD and IDIA as well as Dr Nalini Heeralall-Issur, an Associate Professor of Physics at the University of Mauritius (UoM).
The event took the form of a mini data science school, with a day of lectures from industry and academic speakers followed by a 3-day hackathon. Most of the speakers were Mauritian, with 5 of them having started their professional careers by studying at the University of Mauritius. The call for applications was advertised as widely as possible to allow students and young professionals from a variety of institutions to take part. A lecture theatre at the Reduit campus of UoM was booked for the first day, with the hackathon taking place in the Core Building on the Ebene campus.
Big Data Mauritius was opened on Tues June 21st by Prof Sanjeev K Sobhee, the UoM Acting Vice-Chancellor. He spoke about the remit of DARA Big Data and also covered the DARA Project, who have sponsored several African students to study astronomy at UoM. Prof Anna Scaife’s video presentation ‘What is Data Science’ was shown first, followed by Dr Arrykrishna Mootoovaloo. Dr Mootoovaloo recently completed his PhD research on weak lensing, data compression and Gaussian Processes. He now holds a postdoctoral position at the University of Oxford, using Machine Learning techniques to find unusual objects in the universe. His presentation used fascinating space images to demonstrate an AI-based framework for galaxy classification and anomalous object detection and he finished by taking participants through a live demonstration of Zooniverse.
The next speaker was Ari Ramkilowan from South Africa, a Natural Language Processing (NLP) Research Scientist at Helm Africa. His presentation ‘A Whirlwind Tour of NLP’ took participants through the roots of NLP and its many applications, also taking a look at future usage. He spoke about his work creating chatbots and shared a comprehensive list of learning resources that he has found useful. Dr Zafiirah Hosenie then gave her talk ‘From Astronomy to Mixed Reality’, where she detailed her career journey to her current position at the Microsoft Mixed Reality and AI Lab in Cambridge. She talked about the futuristic work she is involved with at Microsoft, with projects including high-fidelity neural rendering for 3D object creation and the use of synthetic data to create realistic human faces and expressions.
After lunch it was the turn of David Bastien, who spoke about returning to Mauritius following his studies and founding a start-up company, Hydrus Labs. Hydrus aims to solve a range of issues for businesses across Mauritius using research, methodologies and data-driven solutions. David presented a range of data analytics tools and showed some of his MSc research work where he used machine learning to create simulated radio galaxies. He took participants through some of the projects that Hydrus has been involved with, such as analysing tourism data to provide useful insights to local hotels and creating the first ever dataset of the Mauritian Creole language. David has many more projects lined up and has already provided employment for several Mauritians through his growing company.
Following David was Suhaylah Peeraullee of Mauritian-based StraConsult consulting firm, who works on the insightful AfroBarometer project. AfroBarometer surveys countries across Africa to get data to use in forming policies that are responsive to national concerns. Suhaylah spoke about the work that StraConsult does, which includes analysing the data gathered for Mauritius. She covered the methodology behind the surveys and gave examples of how results are used, showing selected findings and corresponding actions from recent years. StraConsult founder Amedee Darga was also in attendance and he added his thoughts on data usage to improve society at the end of Suhaylah’s talk. Dr Nadeem Oozeer, a data scientist at the South African Radio Astronomy Observatory (SARAO) in Cape Town, then joined via Zoom to give a presentation on astronomical Machine Learning applications. Nadeem has worked closely with DARA Big Data and partners to put together several data science learning events, such as Big Data Africa and the Data Science Intensive programme. He is well versed in research and also in using data science from an industry perspective and spoke at length about his experiences.
To end the day Dr Rajin Ramphul gave a video presentation on data science in industry. Rajin has a research background but now works at OrderIn in South Africa as a data scientist. He spoke about his journey through academia into industry and spoke about how companies use data for various ends, such as working out inventory requirements and forecasting demand. He also covered how data science is approached very differently in academia and industry and the best deployment of machine learning systems.
On Wednesday the hackathon began; Dr Nikhita Madhanpall of the OAD opened the day with an explanation of what lay ahead for participants. The 27 participants were put into 6 teams that balanced gender, Python skill levels and educational institutions. Everyone worked through several tutorials during the day using the cloud computing platform provided by IDIA. On Thursday everyone began work on the hackathon task, which was to use data scraped from Twitter to analyse sentiment on a COVID19-related topic of the teams choosing. Tutors Nikhita, Ari, Zafiirah, Arrykrishna and David were available at all times to give support and advice. In the afternoon participants gathered for a ‘fireside chat’, an informal group conversation covering any aspect of careers, data science or anything else that participants found useful to discuss. Topics included industry vs academia, tips on applications and interviews, how to network and women in STEM.
On the final day participants started work on their presentations, analysing data and rehearsing the content of their talks. All 6 teams did an excellent job with their presentations and showed a good grasp of key concepts such as NLP, machine learning and dataset preparation. There was a good range of subject matter, looking at public attitudes to mask wearing, weight gained during lockdown, mental health and global herd immunity. While it was hard to choose the winning teams it was decided that Team Incredible Analysts should be given second place with their look at changes in the education system resulting from COVID19. Team Undefined were declared the overall winners with their excellent assessment of public reaction to vaccine authorisation for young children. There was also a special prize for Love Achondo of Team 4MTGT, who was declared the best overall presenter.
It’s obvious that there is a lot of data science talent in Mauritius; participants were dedicated, engaged and showed a lot of promise and enthusiasm. Dr Madhanpall said ‘I found that participants were not only eager to learn but determined to do their very best from day one of the hackathon. It was a pleasure to work with such driven and creative individuals.’ Feedback received after the event showed that those who took part found it a very valuable learning experience, some comments are shown below;
‘From the use of data science in physics to the entry of Machine Learning technologies in industry, every talk covered different aspects of data science, from the basics to advanced applications. It was a very informative experience overall and difficult to single out the best talk’
‘Ex-students from the University of Mauritius came and shared their experience as presenters and tutors, which is encouraging for future batches of students who will come in the post graduate world’
[The most memorable thing was] ‘working together with a team I never have before, in synergy and enthusiasm, to produce a remarkable output we were all proud of’
‘All I want to say is that after this four-day workshop I am walking away with a bag full of memories, knowledge and experience. Thank you for that’