Analyzing the Determinants of Guest Experience in Istanbul's Airbnb Market: An Advanced Topic Modeling Approach via Latent Dirichlet Allocation

Abstract

This study embraces an inductive approach to comprehensively examine the various factors influencing customer experience and satisfaction in Istanbul's rapidly evolving accommodation-sharing economy, with a specific focus on Airbnb. The research undertakes an extensive analysis of a substantial dataset comprising 508,746 Airbnb reviews collected from Istanbul, marking a significant endeavor in understanding the nuances of customer preferences and expectations in this domain. The process begins with a thorough preprocessing of the textual data, ensuring clarity and relevance in the information analyzed. Following this, the study employs Latent Dirichlet Allocation (LDA), a sophisticated statistical model, to identify and extract 32 distinct topics from the user-generated content. These topics, embedded within the reviews, provide a rich source of insights into the guest experience. The extracted topics are systematically categorized into several key dimensions, offering a structured framework for analysis. These dimensions include detailed assessments made by guests, locational attributes of the accommodations that range from central urban areas to more peripheral locations, and both the tangible and intangible aspects of the Airbnb listings. Additionally, the study examines the management practices of the hosts and the overall quality of service, factors that are crucial in shaping guest satisfaction. Each of these dimensions offers a lens through which the intricate aspects of customer experience in the shared accommodation sector can be understood and evaluated. To explore deeper into the intricate relationships among these topics, the study employs hierarchical Ward Clustering. This statistical technique is instrumental in revealing the complex interplay and subtle connections between the various topics. Such an approach is pivotal in elucidating the multifaceted nature of customer experience in the peer-to-peer accommodation context. The analysis aims to provide a comprehensive and layered understanding of the determinants that shape guest experiences in Istanbul's Airbnb sector. By offering a detailed, multi-faceted perspective on the drivers of customer satisfaction, this study contributes significantly to the body of knowledge in the field, enhancing the understanding of key factors that influence guest experiences and satisfaction in the dynamic and diverse landscape of Istanbul's accommodation-sharing economy.