ANNUAL CONGRESS 2025

Promoting Sustainable Heritage Tourism Through Measuring Tourists' Perception Based on Machine Learning

Not scheduled

20m

Yildiz Technical University, Istanbul

Poster Track 06 | URBAN CULTURES AND LIVED HERITAGE

Speaker

Dr JINGCHAO WANG (School of Architecture and Materials, Chongqing Polytechnic University of Electronic Engineering)

Description

Heritage tourism plays a pivotal role in balancing cultural preservation and urban development, particularly in regions where historical significance and modern expansion intersect. As urbanization accelerates globally, heritage sites face threats such as commodification, gentrification, and environmental degradation, challenging the sustainable management of cultural landscapes. Existing studies on heritage tourism have primarily focused on qualitative methods, including surveys and interviews, which are constrained by limited sample sizes, subjective biases, and high implementation costs. Moreover, these approaches often fail to fully capture the complex, multidimensional interactions between tourists and heritage sites, leaving significant gaps in understanding how perceptions evolve in response to spatial, emotional, and cultural factors.

To address these limitations, this study employs advanced machine learning techniques and multi-source data analysis to construct a comprehensive tourist perception measurement model. This model integrates spatiotemporal analysis, natural language processing, and computer vision to provide a holistic framework for evaluating tourist perceptions in heritage tourism. The research leverages data from online travel platforms, including 18,380 textual reviews and 30,357 images, to analyze tourist perceptions across six dimensions: spatial structure distribution, heritage type clustering, sentiment orientation, typical opinion extraction, image content recognition, and historical image.

The empirical study focuses on Zhoushan, China, a historic and cultural archipelago renowned for its Buddhist temples and maritime landscapes. Spatial analysis using GIS tools reveals significant clustering of tourist activity around Putuo Mountain and Zhujiajian, while historically rich areas like Dinghai Ancient City are underutilized. Sentiment analysis, conducted through natural language processing, identifies generally positive perceptions of heritage sites but highlights challenges related to accessibility and underdeveloped infrastructure in less frequented areas. Image content recognition, utilizing computer vision algorithms, reveals a dominance of natural landscapes and Buddhist culture in tourist-generated imagery, with limited representation of urban heritage and local traditions.

The findings underscore key challenges in Zhoushan’s heritage tourism development, including perception misalignment, imbalanced resource utilization, and a lack of interconnectivity among heritage sites. For instance, while Buddhist and natural landscapes dominate tourist narratives, urban and folk heritage remains underrepresented, leading to an incomplete depiction of Zhoushan’s cultural identity. Additionally, transportation bottlenecks and fragmented resource management hinder the integration of diverse heritage assets into a cohesive tourism framework.

By addressing these challenges, the study offers several contributions. First, it demonstrates the potential of integrating machine learning techniques with open-source data to overcome limitations of traditional qualitative methods, enabling large-scale, high-precision analysis of tourist perceptions. Second, it provides actionable insights for policymakers and stakeholders, advocating for a more balanced and inclusive approach to heritage conservation and tourism planning. Recommendations include diversifying narratives to incorporate underrepresented heritage elements, improving infrastructure to enhance site accessibility, and fostering collaboration across administrative boundaries to achieve cohesive resource management.

This study advances the field of heritage tourism by introducing a replicable, data-driven methodology for measuring and analyzing tourist perceptions. The proposed framework bridges the gap between traditional qualitative studies and emerging digital tools, offering a robust foundation for future studies on sustainable heritage tourism. The findings contribute not only to the theoretical understanding of tourist perceptions but also to the practical management of heritage resources, paving the way for more resilient and inclusive cultural landscapes.