Data-Driven Video Art is a form of generative video art in which the visual and temporal composition is algorithmically determined by external data sources or dynamic datasets, distinguishing it from rule-based or manually authored approaches by prioritizing data as the primary generative agent.