A comparative study of user- and organization-generated hotel imagery across diverse geographical environments

Visual imagery greatly affects impressions of hotels and their geographical context (i.e. the properties themselves and their surroundings). This study compares how hotel zones are depicted in different destinations. It specifically considers cognitive, affective, and overall aspects along with several stakeholder groups, namely tourists, organizations, and professionals. Deep learning techniques were employed to analyze user-generated photos (UGPs), organization-generated photos (OGPs), and professional-generated photos (PGPs) in three popular American cities. Google Vision API, DeepSentiBank with VaderSentiment, k-means clustering, and aesthetics automation with analysis of variance were, respectively, used to detect the objects, sentiments, colors, and aesthetics contained in photos. A dataset was assembled, comprising 7,953 images from TripAdvisor, to examine the interplay between content creators’ visual representations of hotels across geographical settings. The results revealed substantial perception gaps between customers and sellers. Differences also emerged in how UGPs and OGPs/PGPs showcase geographical features. These findings add nuance to the academic discussion on brand image by exploring various content creators and hotel zones. Strategic recommendations are provided on using hotel images to boost destination appeal, guide hotel marketing campaigns, and optimize platform design.