Deconstructing and Reconstructing a Periodical Archive by Tracking, Tagging, and Extracting Images of Faces
Symposium:
Session Title:
- Data Visualization
Presentation Title:
- Deconstructing and Reconstructing a Periodical Archive by Tracking, Tagging, and Extracting Images of Faces
Presenter(s):
Venue(s):
Abstract:
This talk presents progress in an ongoing data-mining and visualization project exploring the archives of Time magazine. Our goals are to extract and re-contextualize images of faces (along with their associated metadata) from the entire Time magazine archive (dating back to 1923), and to create public-facing web-based visualizations (as well as interactive VR and AR installations) that display our results and invite viewers to explore how representations of human faces have changed over time within the archive. This work combines work in machine learning, data analytics, interactive design, and the humanities, bringing together researchers from across disciplines. A parallel project objective is to develop an interdisciplinary methodology for systematically deconstructing any periodical image archive.