From data to discovery: the role of economists in the era of machine learning

The advancement and refinement of machine learning techniques marks a thrilling era for scientists engaged in the study of human behaviour through data. Combined with the growth in computational power, the rapid accumulation of behavioural data, and breakthroughs in neuroscience, this paves the way for constructing models that can more accurately understand, predict and shape human behaviour. Machine learning excels at detecting nuanced patterns in data that would typically elude human analysts and could support the development of more refined models. Nonetheless, domain knowledge remains indispensable when working with real-world data, which are characterised by issues such as non-random sampling, statistical relationships that are not indicative of causality, and measurement error. When overlooked, these issues can result in biased estimates, misleading narratives and poor decision-making. In this lecture, Prof Rulof Burger draws on examples from his own research to explain how such biases arise, and how the tools and theories developed in economics can provide reliable answers to important questions.

WATCH THE INAUGURAL LECTURE HERE

Short biography

Rulof Burger is a professor of Economics at Stellenbosch University (SU), where he teaches Microeconomics, Econometrics and Behavioural Economics. His research integrates data and economic theory to explore human behaviour, often in the fields of labour, education and health. Having completed a master’s degree at SU, he joined the University’s academic staff in 2002. He went on to obtain another master’s degree from Cambridge University in 2004, as well as and his doctorate from Oxford University in 2012. Currently, he forms part of the Research on Socio-Economic Policy (ReSEP) group at SU and is one of two South African professors affiliated with the Abdul Latif Jameel Poverty Action Lab (JPAL). Burger is also a co-founder of Predictive Insights, a data science company that performs behaviour modelling for various South African and international entities.

RSVP HERE