A machine-learning approach to examining users’ responses to travel content on YouTube

This study develops an integrated machine-learning framework to analyse user responses to travel content on YouTube. Using 10,893 comments from Expedia’s YouTube channel, we combined Latent Dirichlet Allocation with topic modelling, sentiment analysis, and K-means clustering. Six distinct topics emerged: travel vlog enjoyment, tourist attractions, recommended destination appreciation, travel wishes list, content production quality, and the city-living experience. Sentiment analysis revealed complex emotional patterns with a predominance of joy, trust, and anticipation. User clustering identified distinct viewer segments, whereas Shapley Additive Explanations analysis showed that travel vlogs generated the highest engagement. As a limitation, this study relied on text-mining analysis of user comments from a single channel without incorporating the visual and auditory elements of the video content. Nevertheless, the findings can help tourism marketers create targeted content for specific viewer segments, prioritise travel vlog formats, and develop sentiment-aware recommendation systems that enhance user experience and social media engagement.