BERT-Based Temporal Sentiment Analysis of Saudi Tourism Events

Abstract

This study applies a BERT-based temporal sentiment analysis framework to examine English-language social media discourse surrounding major tourism and sporting events hosted in Saudi Arabia. A dataset of 10,000 tweets collected between December 2023 and December 2024, including discussions related to the FIFA Club World Cup, Formula 1 Saudi Arabian Grand Prix, and other major Saudi-hosted sporting events, was analyzed to investigate sentiment dynamics across pre-event, during-event, and post- event stages. The fine-tuned BERT model achieved an F1-score of 0.85, outperforming traditional machine learning and deep learning baselines. Results reveal strong positive sentiment during event periods, with measurable variations across different temporal stages. The findings demonstrate the value of transformer-based sentiment analysis for understanding international audience engagement with Saudi mega-events and provide empirical insights relevant to tourism development and Vision 2030 objectives.