Digital threads of architectural heritage: navigating tourism destination image through social media reviews and machine learning insights

Abstract

While emphasizing sustainable development goals for cultural heritage preservation, tourism destination image based on the predefined traditional survey often fails to meet the dynamic tourist demands. Taking the Nanjing Museum of Modern History in China as a case study, this paper innovatively combined the Latent Dirichlet Allocation (LDA) and Importance-Performance Analysis (IPA) to quantitatively extract and assess the destination image from the social media data. The results demonstrated that the tourists mainly focused on 12 topics from individual, environmental, and social backgrounds, mostly on architectural style, transportation, and location. Tourists perceived most positive sentiments in historical culture (84.94%), and architectural style (70.54%), but most negative sentiments in ticket purchasing experience (27.60%), living and service facilities (14.09%). The IPA results reveal that historical culture is the strength, yet there is an overemphasis on historical events. This research marked a methodological advancement by combining LDA and IPA, enabling a detailed analysis of vast online reviews and providing critical insights for improving the tourism destination image in architectural heritage studies.

Keywords

• destination image
• machine learning
• big data
• smart city
• built environment