Abstract
This paper describes the first comparative multimodal dataset focusing on climate change communication across two major social video platforms prevalent in different parts of the world: BiliBili (China) and YouTube (International platform but here we focus on the audience based in United Kingdom). These platforms were selected because their long-form video formats lend themselves particularly well to narrative analysis, offering advantages over other forms of social media. The dataset comprises video cover images, metadata, engagement statistics, and textual content (titles, descriptions, and transcripts) for 2,798 videos (2,092 from BiliBili and 706 from YouTube), ranging from May 2007 to April 2025. All data are hosted on Mendeley Data to support Open Access and reuse. The dataset offers substantial potential for researchers in environmental communication, digital humanities, and computational social science, enabling cross-cultural analysis of public discourse and storytelling practices related to the climate crisis.