Collections as data
Also: collections-as-data
The movement to make library, archive and museum collections usable for computational research — as text, data and metadata — not only as items to view one at a time.
Last updated
A digitised collection is not necessarily usable as data. A run of periodicals scanned as page images can be viewed, one page at a time, but not searched, counted, compared or read at scale. The collections-as-data movement — set out in the Santa Barbara Statement of 2017 — argues that cultural heritage institutions should make their holdings available for computational use as well as for viewing, with the documentation, licences and provenance that responsible use requires.
Why it matters for your research. Distant reading depends on it: without text, there is nothing to search, group or trace. And the gap is uneven. English-language collections in well-funded institutions are far more likely to be available as data than vernacular collections held in the Global South, many of which exist online only as images. What can be read at scale shapes what gets studied.
In Archēglyph. Turning page images into text is the first thing Archēglyph does: each page is read part by part, every passage keeps its place on the page image, and a collection can be downloaded whole as an edition in open formats. Reading open repositories’ collections where they already live is on the project roadmap as an aim.
Not to be confused with. Digitisation, which produces the images; collections as data is the step after it. Nor with “big data” — a collection of a few hundred pages is data too, if it can be computed over.