Original Research

Employing sentiment analysis for gauging perceptions of minorities in multicultural societies: An analysis of Twitter feeds on the Afrikaner community of Orania in South Africa

Eduan Kotzé, Burgert Senekal
The Journal for Transdisciplinary Research in Southern Africa | Vol 14, No 1 | a564 | DOI: https://doi.org/10.4102/td.v14i1.564 | © 2018 Eduan Kotzé | This work is licensed under CC Attribution 4.0
Submitted: 19 April 2018 | Published: 15 November 2018

About the author(s)

Eduan Kotzé, Department of Computer Science and Informatics, University of the Free State, South Africa
Burgert Senekal, Unit for Language Facilitation and Empowerment, University of the Free State, South Africa


South Africa is well known as a country characterised by racial and ethnic divisions, particularly for the divisions and conflicts between the white population and black population. This study uses the Twitter platform to analyse the discourse around the controversial town of Orania, a minority Afrikaner community that aims to preserve their Afrikaner culture. In doing so, we make use of sentiment analysis, a subfield of natural language processing (NLP). We follow a lexicon-based approach using four different publicly available data sets to test how the discourse around this minority community can be analysed. We show, based on the discourse on Orania on Twitter, that (1) Orania is mostly depicted in a negative light, (2) Orania is mostly seen as a racist community, and (3) Orania is often mentioned in reference to other issues that affect Afrikaners directly, such as farm attacks, first language education and land expropriation without compensation. Our study also shows that using lexicons as a sentiment analysis technique was not sufficient in the automatic detection of abusive language, but rather the sentiment of the tweet. Suggestions are made for further research that focuses on the automatic detection of abusive language online.


sentiment analysis; Twitter; microblogging; Orania; minority community


Total abstract views: 2583
Total article views: 4590

Crossref Citations

No related citations found.