<![CDATA[This study investigates the potential of online visitor reviews as valid indicators of service quality and sustainability in culinary tourism destinations, focusing on Semarang Regency, Central Java, Indonesia. Amidst the increasing reliance on digital platforms for travel-related decision-making, the research addresses a critical problem: how user-generated content (UGC) can accurately reflect visitors’ experiences and perceptions, and how this information can be leveraged to support the sustainable development and continuous improvement of culinary tourism. The main objective is to extract key service quality dimensions and sustainability-related elements from online reviews by employing an integrative methodological framework. A mixed-method approach was adopted, combining Natural Language Processing (NLP) specifically sentiment classification using Support Vector Machine (SVM) with qualitative thematic analysis using NVivo software. The data were collected from prominent digital platforms such as Google Reviews, TripAdvisor, and Instagram, representing diverse and large-scale visitor feedback. The analysis revealed that dimensions of service quality namely reliability, responsiveness, assurance, empathy, and tangibles frequently emerged in visitor comments, aligning with the SERVQUAL model. In parallel, sustainability indicators such as the promotion of local cultural identity, environmental cleanliness, waste management, and empowerment of micro, small, and medium enterprises (MSMEs) were also prominent. These findings demonstrate that online reviews provide a rich, real-time, and scalable source of data that can support evidence-based strategies for improving culinary destination services and sustainability. The study contributes both theoretically and practically by offering a big data driven model for tourism evaluation. However, the research also acknowledges limitations in NLP accuracy for local dialects and informal expressions, as well as the digital divide that limits inclusivity. Future studies should refine linguistic models and expand the geographic scope for comparative insights.]]>
