How to “Read” a Destination from Images? Machine Learning and Network Methods for DMOs’ Image Projection and Photo Evaluation

• Published in: Journal of travel research Vol. 61; no. 3; pp. 597 - 619
• Main Authors: He, Zeya, Deng, Ning, Li, Xiang (Robert), Gu, Huimin
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01.03.2022; SAGE PUBLICATIONS, INC
• Subjects: Machine learning; Marketing; Projected images; Tourism; Tourist attractions; User generated content; destination image; network analysis; core–periphery structure; machine learning
• ISSN: 0047-2875; 1552-6763
• DOI: 10.1177/0047287521995134

Online photos can reflect tourists’ received destination image and be used to project destination image by destination marketing organizations (DMOs). Studies have identified a gap between projected and received images, highlighting the difficulty DMOs face when selecting content to project the “right” image. Taking an audience-driven perspective, this study analyzed information from user-generated content (UGC) to guide the selection of organization-generated content (OGC) on social media. Using a machine learning algorithm, we extracted connected cognitive and affective elements of received and projected images from UGC and OGC. The elements and their relationships retrieved from UGC were then used to construct a semantic network. The network informs the core–periphery structural information of each element and guides DMOs’ image projection and content selection. Studies with two independent samples demonstrated that an OGC photo whose projected images matched consumers’ central impressions, particularly affective ones, could induce higher online engagement.