We should all be Data Scientists: Why SOCIAL and SOCIETY are the missing links.

This post originally appeared on the UP EBIT Website

Dr Vukosi Marivate, the University of Pretoria (UP) ABSA Chair of Data Science, is based in the Department of Computer Science at the University. With all the talk about the 4th Industrial Revolution (4IR), Dr Marivate believes we need a more nuanced view of how the emerging technologies connected with the 4IR will impact all of our lives. This post covers two talks that Dr Marivate gave: one as a public lecture at North-West University, and another as a seminar at the Centre for the Advancement of Scholarship at the University of Pretoria.

We can be optimistic that technologies will make a positive contribution to the ways we live, to our health, our education and our overall well-being. At the same time, we need to understand how such technologies augment and change our society. Through data science, one can look at the world through data collection, manipulation and modelling. The data, more often than not, is about people. As such, society has to be part of the conversation when we talk about data science, machine learning and artificial intelligence.

As data scientists, we use machine learning to identify patterns in data. Most of the data used by Dr Marivate and his group, is from text. Text data is at the same time abundant and scarce, especially for languages that are underrepresented or from areas where it is hard to collect data formally, even though a language is widely used. Therefore, when building decision-making models with this data, one has to be careful not to discriminate.

This is where the social, as in our social sciences and humanities, holds great promise. The same can be said in general when we use data science for society. Data is ultimately about people. How do we make sure that we represent our values and understand that data is not biased, while providing decision-making tools for society? We have an opportunity to champion more interdisciplinary research in the Faculty of Engineering, Built Environment and Information Technology Engineering, Built Environment Information Technology (EBIT), while also working with other faculties to bring data skills and computational thinking to the whole University. Breakthroughs are now more and more driven by diverse teams working across disciplines.

The work done by the Data Science for Social Impact Research Group transcends disciplines and looks at different ways to approach social challenges and use data science as a possible way to understand ourselves and the world more fully. To do this well, we work not only practise science of a high standard, but also work to understand how we can give back to the community we live and work in. Past work by the group investigated education, public safety, and energy. For natural language processing, the group is currently examining ways to leverage text as an information source for tasks such as election understanding, robustness and cyber safety.