From Text Mining to Visual Classification

Focusing on the specific case of cinema owner and film distributor Jean Desmet’s digitised business archive, this article discusses how computational approaches may facilitate the archive’s unlocking for researchers in the Dutch national research infrastructure – CLARIAH –Media Suite. To this end, the article considers previous computational approaches to film- related sources in New Cinema History research in a historical perspective, suggesting a novel approach which combines text mining and visual classification. The article argues that such a combination is necessary to yield results which reflect the archive’s material heterogeneity and complexity, and that it offers a new direction for computational approaches in New Cinema History and their conceptualisations of film-related materials as historical sources.