Topic modelling of festival-visitors' opinions: A case study of Sziget Festival

Abstract

The Sziget Festival is one of Europe’s biggest music events, held every August on the Óbuda Island in Budapest. The elements of festival experiences were analysed through social media opinions of Sziget Festival attendees published between 1 January 2014 and 31 July 2022, using structural topic model (STM) computer algorithm and latent Dirichlet allocation (LDA). Our findings support existing literature, showing that festival experiences are cognitive-affective in nature with key elements of socialisation, environment, and festival value, some of which are inter-connected. Although this kind of research is rare in tourism, our paper argues for its usefulness and provides an example of application, using consumer data to facilitate decision-making and support the festival industry. The paper concludes with recommendations for future research, as well as incorporating LDA with other qualitative and quantitative research.