How can we better understand sustainability transitions when traditional data often fail to capture their speed, complexity, and local specificities? This chapter argues that novel, non-traditional data provide a crucial complement to statistics by offering granular, timely, and context-sensitive insights into environmental and socio-economic change. It makes five contributions. First, it uses Large Language Models to map the growing use of novel data sources – e.g., search queries, social media traces, satellite imagery, citizen-generated information – across multiple sustainability domains. Second, it presents a Google Trends case study showing how online search behaviour reveals temporal, spatial, and thematic patterns in public attention to sustainability across EU countries. Third, the chapter examines validation and generalizability challenges, arguing for the integration of non-traditional data with high-quality conventional sources. Fourth, it reflects on key data governance issues in sustainability research. Fifth, it outlines strategic priorities for policymakers to effectively leverage these data for sustainability transitions.