hide
Free keywords:
Energy transitions, Coal, South Africa, Discursive power, Topic modeling, Text as data
Abstract:
This paper investigates media discourses on coal in South Africa to understand the state of progress of its energy transition. We approach discourse as a key element in policy sequencing, shaping the conditions under which regulatory change becomes possible. Using an integrated text-as-data pipeline, including topic modeling, named entity recognition, and sentiment analysis, we analyze approximately 8,000 national newspaper articles from 2010 to 2024. Topic modeling reveals five discourse clusters: Mining industry, Transition politics, Energy crisis, Mining affected communities, and Politics in mining and energy. While new technologies receive attention – particularly around the announcement of the Just Energy Transition Partnership with South Africa in 2021 – explicit discourse on coal phase-out remains marginal. At the same time, we highlight that these technologies are frequently framed as substitutes for structural change. Positive sentiment toward the mining sector and transition politics suggests that the transition is not framed as a coal exit. In this sense, the observed framing reveals not only media priorities but also the discursive power of dominant actors in structuring transition pathways. Our findings underscore the value of combining computational and qualitative approaches innovatively to examine socio-political transition discourses, with implications for broader applications of text-as-data in policy research. Our results inform debates on the political economy of energy transitions at the interplay of international climate finance.
