Every researcher knows that working together across academic disciplines is notoriously hard.
Even seemingly related scientific fields – such as computer science and information technologies – are often divided by barriers such as different languages, academic cultures and approaches to problem solving. And it does not get better when we consider fields as disparate as astrophysics and social psychology.
UKZN’s Big Data for Science and Society (BDSS) collaboration is an ambitious initiative that aims to break down these barriers. It unites a number of traditionally unrelated disciplines, including social science, social psychology, physics, computer science and geospatial science, under the umbrella of big data. Led by Principle Investigators, Professor Kavilan Moodley (Astrophysics – College of Agriculture, Engineering and Science), Professor Maheshvari Naidu (Social Science – College of Humanities) and Professor Onimiso Mutanga (Earth Observation – College of Agriculture, Engineering and Science), it is based on the simple observation that staggering developments in data analysis, machine learning and 4IR technologies offer huge potential for many academic disciplines and that they can serve as a common purpose under which exceptional transdisciplinary projects can prosper.
In this spirit, collaborators from across the Colleges of UKZN and 30 researchers and students from many parts of the world, particularly the African continent, came together for the first BDSS workshop on KwaZulu-Natal’s north coast on 13 and 14 February. The first day was dedicated to sharing and advancing progress on 12 projects running under the collaboration. For example, a team from computer science and social psychology is using reinforcement learning simulations to model and study interactions in groups.
Researchers from geospatial science and social science want to use remote sensing techniques to monitor floods in informal settlements, and food security in rural areas. Astrophysicists, together with machine learning experts, are developing new methods to analyse experimental data from South Africa’s radio telescopes, while quantum physicists use intelligent data mining techniques to secure communication.
The stimulating two day workshop was put together by Dr Maria Schuld who is also one of the co investigators or CoIs in the project. The data skills workshop on the second day was facilitated by four of Durban’s Data Carpentries instructors. Starting from zero programming skills, the workshop slowly guided participants to a level where they were able to import, slice and clean data in Python’s Pandas library.
The BDSS workshop created a feeling of excitement among the participants, who realised that they are involved in something much bigger than their own discipline in transcending trodden paths of research to push answers to our most burning questions one step further.
Photograph: Andile Ndlovu