Building the Knowledge Graph of Cinema
Every film creates more than moving images. Behind the finished work lies a complex history of people, places, decisions, creative experiments and practical discoveries. Some of it survives in credits and production documents. Much of it remains scattered across disconnected systems or disappears into personal memory.
On 24 September 2026, Filmster Network founder Anush Venkataraman took this problem to Big Data LDN 2026, presenting Building the Knowledge Graph of Cinema: Turning Film into Data Infrastructure. The session brought filmmaking into a conversation usually dominated by enterprise data and engineering, asking a deceptively simple question:
What if cinema could learn from its own accumulated experience?
Every film is a network
A film is made through relationships. Directors collaborate with cinematographers, production designers, actors and technicians. Scripts become scenes; scenes require locations, schedules and resources. Every decision influences another. The industry records fragments of these connections. Databases list credits; production software manages schedules and documents; archives preserve films. Yet the relationships between these records often remain difficult to trace.
A knowledge graph offers a way to connect them. People, projects, locations, creative artefacts and production activities become identifiable entities linked by meaningful relationships. Natural language relation extraction can help identify connections in documents, while multimodal machine learning can reveal patterns across text and imagery.
Imagine researching a location and discovering the productions filmed there, the crews involved, the practical difficulties encountered and the techniques employed. These connections could help future teams investigate relevant experience and build upon existing knowledge. The challenge is to make these relationships trustworthy. Context, provenance, verification and consent matter as much as the technology that connects them.
SO WHAT?
During the presentation, a slide interrupted the technical discussion with two words: SO WHAT? The audience laughed. It was precisely the question the technology needed to answer.
A filmmaker may use connected knowledge to discover collaborators through verified work. A producer may learn from previous scheduling challenges. A film school may teach production practice through real-world examples. Researchers may explore how techniques and creative relationships evolve over time.
These applications require further development and testing. Their value depends on whether they help people make better-informed decisions.
And there is a boundary worth protecting. Patterns can illuminate past choices, but filmmaking depends on imagination, risk and human judgement. A model may identify what has happened before; the filmmaker decides what ought to happen next.
Patterns should remain knowledge, not become instructions for creativity.
An evolving memory of cinema
The larger ambition is a collective memory of filmmaking. Each production contributes experience; future productions can build upon it, question it and add to it. Over time, connected knowledge could support creative work, professional recognition, education and research. It might also reveal links between cinema and other fields, including tourism, architecture and cultural history.
At Big Data LDN, the full room, engaged audience and questions following the session demonstrated an appetite for this conversation across disciplines. For Filmster Network, the presentation marked another step in exploring how cinema's fragmented information can become useful, responsible knowledge infrastructure.
Cinema has spent more than a century accumulating creative knowledge. How much of it have we actually learned to preserve?
Download session presentation, with notes.
Building the Knowledge Graph of Cinema: Turning Film into Data Infrastructure was presented by Anush Venkataraman, founder of Filmster Network, at Big Data LDN 2026 on 24 September, from 14:40 to 15:10, at the AI Context Theatre, Olympia London.
Filmster Network thanks Big Data LDN for selecting the session, everyone who attended and contributed to the discussion, and co-founder Anjana Krishnan for her support throughout the event.