Du Bois at the 1900 Paris Exposition
The hand-drawn statistical charts and the photographs assembled for what the Library of Congress catalogues as the "Negro Exhibit of the American Section at the Paris Exposition Universelle in 1900".
Whose Du Bois this is
This is not the W. E. B. Du Bois Papers held by UMass Amherst Special Collections. Those are
served through credo.library.umass.edu, which answers every scripted request with a bot check and
was not harvested.
The UMass finding aid for those papers is reachable and is held separately as
umass/dubois-papers.
Source
loc.gov's public JSON API, collection african-american-photographs-1900-paris-exposition. No
key. It catalogues every sheet individually and carries the things a viewer needs: a summary line
saying what each diagram shows, subject headings, the medium and board size, and JPEG derivatives
at several widths.
It rate-limits hard. Two requests in flight is enough to earn a 429, and once you have earned
one the block persists for a while rather than clearing on the next request. The scraper runs one
request at a time with a 1.8 second pause, which takes about twenty minutes for the full
collection, and caches every item's JSON under input/loc-cache/ so a second pass costs nothing.
Record types
| Type | Notes |
|---|---|
Plate |
A chart. Classified by genre (Diagrams, Drawings) or by an ink-and-watercolour medium statement. |
Photograph |
Everything else in the collection. |
Series |
Derived from the bracketed prefix on plate titles, with a context note written for this collection. |
Notes
- Series membership comes from the title. The Library prefixes a caption with its study in
square brackets, as in
[The Georgia Negro] Illiteracy.A bracketed title with nothing after it is a cataloguer-supplied title for an untitled photograph, not a series, and is treated as such. - No images are copied. Records carry URLs into the Library's own image service.
- Not every plate has a description. Sheets without one keep no
summaryrather than being given an invented caption. - Subject headings are the Library's, including the period language they preserve.
Run
pnpm scrape umass/dubois-paris-1900 # about 20 minutes cold, seconds warm
pnpm push umass/dubois-paris-1900 --env dev