From pixels to perceptions: analyzing rural tourism image diffusion through Douyin-style short videos

Purpose By integrating text mining, sentiment analysis and visual analysis, a comprehensive framework of destination image formation and diffusion is developed to explore the impact of short videos on rural tourism destination image construction and diffusion in the digital economy era. Design/methodology/approach Based on the cognitive-affective model, this study examined the destination image projected by 16,924 short videos related to rural tourism destinations from Douyin. The Latent Dirichlet Allocation model is employed for the classification of cognitive attributes, the SnowNLP library is utilized for the analysis of the emotional tendency of destination images and the deep learning model Dover is employed for the extraction of visual index information including video aesthetic quality and technical quality. Additionally, drawing on sensory marketing theory, the ordinary least squares model is employed to investigate the primary factors predicting the diffusion effect of destination image. Findings The findings demonstrate that video aesthetic quality and thematic content focused on travel experiences and culinary attractions significantly enhance diffusion effectiveness. Furthermore, shorter video formats (16–60 seconds) maintain superior image diffusion efficiency compared to longer formats, while higher follower counts and official account verification status positively relate to diffusion outcomes. Originality/value This research reveals the formation of rural tourism destination image and its diffusion effect in the context of short video platforms, and offers evidence-based strategies for destination management organizations to optimize short video marketing efforts.