the repository is still empty. add the first item under info → upload.
This research brings together two types of material:
(i) Collected data comes from direct observation and engagement: photographs, audio recordings, interviews, videos, field notes, and news.
(ii) Scraped data is extracted from digital sources and existing information systems.
The interface does not provide a complete or definitive reading. It creates the conditions for examining connections between fragments.
Developed within the Studio Caruso at ETH Zürich, this research focuses on the Schlachthof Zürich as the subject of the semester.
contributors: Ava Quiblier, Basil Menzi, David Biedermann, Kishon Umamageswaran, Livia Walther, Matvej Golovatyuk
Every item needs its "vector" (meaning numbers) so the similar search can find it. New uploads get it automatically. Items uploaded earlier need this one step. Photos, GIFs and videos are described by the AI first. Keep this window open until it is finished.
Add many text items at once. Every row of the spreadsheet becomes one text item in the repository.
Add a regulation document (laws, guidelines, rules) as a .txt file. The page finds its entries, for example "• Art. 15 (title): text". Every entry you tick becomes its own text item in the repository. Clicking it shows the whole document with the entry highlighted.
Entries found. Tick the ones that should appear in the repository. "change keywords" sets other keywords for one entry.
Take a photo. An AI looks at it, picks the matching keywords and words for what is visible, and shows the related media like a search. The words found in the texts are highlighted. The photo is only analysed, it is not saved.