From Narratives of Value to Valuable Algorithms

Over the past decades, many collections of moving images have been digitised, catalogued and preserved in various formats, at different scales, and in different archives around the world. The BFI National Archive is one of the most important collections of film and television in the world. It holds a living and expanding record of the evolution of the moving image over the past 90 years, including digital-born content across major television channels in the UK. Since the 1980s, the BFI archive has been capturing and preserving off-air television from UK public service broadcasters, including all advertisements broadcast across all main commercial PSB channels since 2016. Advertisements are the “literature of economics”, in aggregate these TV adverts are likely to be the largest collection of its kind worldwide and one of the most significant records of consumer culture in the UK over the past decade.
However, this collection is fragmented and scattered across thousands of hours of digital video, and despite its cultural significance it remains entirely undocumented and therefore inaccessible. This project aims to:

Define this wealth of material into a usable collection.
Develop a systematic understanding of the collection as a record of consumer culture in the UK.
Create the kind of documentation that opens this resource as a public record for research across disciplines and to wider audiences. 

The challenge to bring such collection to light is technical as much as a conceptual ? it requires the student to critically re-imagine and reshape advertisement records as computational objects. To tackle this challenge, this project adopts an interdisciplinary approach that combines interpretive frameworks in film, television and media scholarship with advanced methods in computational humanities to design a “computational lens” with which to machine-see the archive, not only to identify, extract, and catalogue the adverts, but to understand them in context, as an...