Anush Venkataraman

Filmmaking remains one of the world’s largest yet least structured industries, where collaboration, creative decisions, and production history are fragmented across tools and formats. This session explores how modern data techniques can transform cinema into a computable system by converting unstructured artifacts such as credits, scripts, and production documents into structured relational data. Using natural language relation extraction, multimodal machine learning, and knowledge graph architectures, we demonstrate how filmmakers, projects, departments, and visual styles can be mapped into interconnected networks. We further examine how visual embeddings from film imagery enrich these graphs, enabling analysis of both collaboration patterns and cinematic language. The result is a foundation for film intelligence, where creative ecosystems become navigable, analysable, and reusable as structured data infrastructure for the first time at scale.