What Makes Tourism Short Video Engaging? A Machine Learning Perspective

Abstract

Tourism videos wield a significantly greater visual impact on audiences compared to images and text, positioning short videos as pivotal assets in destination marketing strategies. Audiences possess the agency to determine whether to swipe away or engage further with short videos, including actions such as liking, commenting, and sharing, thereby shaping the diverse levels of engagement with tourism content and subsequently influencing travel motivations. In this study, we employ machine vision technology and machine learning models to conduct extensive data analysis on 7770 pure landscape short videos featuring national parks across the United States. Analysis is structured around five dimensions: social, visual, acoustic, textual, travel, and tourism-related, allowing for an in-depth examination of their interplay with engagement metrics. Our findings reveal visitor numbers, the average likes garnered by the publisher’s videos, and a composite variable termed the Push and Pull Matching Index (PPMI) as the foremost factors influencing engagement levels. This research not only furnishes a robust framework for content mining in tourism short videos using big data analytics, but also underscores the efficacy of integrating big data methodologies with established theoretical paradigms.