Data-Driven Approach in Provincial Clustering for Sustainable Tourism Management in Thailand

Abstract

Since tourism industry is one of the most important contributors to Thailands GDP, this study aims to gain insight into the structure of tourism data in Thailand for suitable administration. In machine learning framework, there were 13 predictors from Thailand’s tourism dataset and the average revenue was assigned as response variable which was transformed into multi-level categorical. The
ordinal logistic regression (OLR) was implemented for clustering of 77 Thailands provinces in 8 different scenarios designed by varying the number of clusters to be 2, 3, 4 and 5 together with outlier adjustment technique. Evaluating models’ performance, the numerical results show that the most suitable number of provincial clusters is three and the number of primary, secondary, and tertiary
provinces are 18, 29 and 30, respectively. The significant factors are number of foreign occupancy, Thai visitors, and their average expense. Based on available infrastructures and tourism resources in each cluster, it is challenging for Thailand to recover the foreign tourist arrivals and promote domestic tourism after the COVID-19 pandemic. Through the strategic management in resource allocation and enhancement of marketing efficacy via provincial clustering, this study comprehensively addresses a multifaceted framework of tourism strategies, integrating guidelines, policies, and best practices that span national initiatives, leverage digital marketing, and reinforce soft power. The framework actively involves the participation of small and medium-sized enterprises (SMEs) and aligns with the overarching objectives of sustainable development goals (SDGs), thus fostering a holistic approach to sustainable tourism growth in Thailand.