Comparative analysis of online photos published by managers and customers about ecolodges in Tambopata: a computer vision and text mining approach

This study compares how ecolodges in Tambopata (Peru) are visually represented by customers and managers across two platforms: Instagram and TripAdvisor. We collected manager- and customer-posted photos from both platforms (TripAdvisor: 55 manager, 328 customer images; Instagram: 236 managers, 182 customer images), transformed images into descriptive labels via computer-vision tooling, and applied text-mining techniques. Across platforms and actors, nature/wildlife is the shared core; however, emphases diverge systematically. On Instagram, customers privilege atmospheric encounters and participation, while managers foreground iconic species and architectural identity. On TripAdvisor, customers highlight habitability and nearby landscape views, while managers continue to elevate emblematic fauna and curated infrastructure. These divergences reveal how platform affordances and actor roles shape distinct yet overlapping ecotourism narratives, and the study discusses their theoretical and practical implications.