Opinion Mining using Sentiment Analysis and Topic Modelling for an Entertainment Event

The entertainment sector is one of the critical sectors fostering a vibrant creative economy. Saudi Arabia has recently constructed several entertainment venues in Riyadh to attract visitors and tourists. One such place is Winter Wonderland. Although this event is held yearly, there is not enough information about visitor opinions and satisfaction. Therefore, this paper aims to collect a dataset and analyze the visitors’ sentiments and their discussed topics on Twitter. We employ two methods. The first is classification using three variations of Support Vector Machine for sentiment analysis, and the second is topic modelling using BERTopic. We further examine the discussed topics in light of time and place. The outcome demonstrates that both Linear SVM and RBF outperform Poly SVM. In addition, most of the tweets discuss topics related to tickets, with Saudi cities being the largest source of tweets selling tickets.