Datasetpublic

biglam/index-cards-navy-nurse-corps

US Navy Nurse Corps — Index Cards 25 biographical service cards from the US Navy's Bureau of Medicine and Surgery (BUMED) History Office, one card per nurse, harvested from their Internet Archive collection. Includes members of the "Sacred Twenty" — the first female Navy nurses, appointed in 1908 (Esther V. Hasson, Lena S. Higbee, Elizabeth Hewitt, Della Knight, M. Estelle Hine, Mary Du Bose, Margaret Murray, Sara Cox, Sara Myer, Ada M. Pendleton, Florence T. Milburn, Clare L.… See the full description on the dataset page: https://huggingface.co/datasets/biglam/index-cards-navy-nurse-corps.

sourceHugging Faceotherupdated 4mo agoView on Hugging Face

0likes24downloads

Dataset Card

US Navy Nurse Corps — Index Cards

25 biographical service cards from the US Navy's Bureau of Medicine and Surgery (BUMED) History Office, one card per nurse, harvested from their Internet Archive collection. Includes members of the "Sacred Twenty" — the first female Navy nurses, appointed in 1908 (Esther V. Hasson, Lena S. Higbee, Elizabeth Hewitt, Della Knight, M. Estelle Hine, Mary Du Bose, Margaret Murray, Sara Cox, Sara Myer, Ada M. Pendleton, Florence T. Milburn, Clare L. De Ceu, J. Beatrice Bowman, Elisabeth Leonhardt, Victoria White, Ethel Parsons, Isabele R. Roy) — and later figures including Laura Cobb (WWII POW in the Philippines) and Theresa Burmeister (1918 influenza victim).

Each row pairs the full-resolution card scan with the Internet Archive's ABBYY OCR text and full provenance back to the source IA item.

Source & licence

—Source: US Navy BUMED History Office collection on the Internet Archive.
—Licence: Public Domain. The source cards are works of the United States Federal Government. Every source IA item carries the `Public Domain Mark 1.0`.

Schema

column	type	description
`image`	image	the card scan (JPEG, re-encoded from the IA JP2 derivative)
`ocr_text`	string	full-item OCR text from the IA ABBYY/DjVu output
`page`	int	page index within the source item
`source_id`	string	IA item identifier
`source_title`	string	IA item title
`source_collection`	string	IA collection (`usnavybumedhistoryoffice`)
`source_url`	string	`https://archive.org/details/<source_id>`
`license`	string	`public domain` (per row, all rows)
`theme`	string	`navy_nurse`

How this was built

Part of an exploration of public archival index-card collections for GLAM/AI work. The reusable harvester (resumable per-item parquet shards, polite IA access, JP2 → JPEG decode via Pillow) runs on Hugging Face Jobs against a curated list of IA items.

The productive IA discovery query (generic full-text is too noisy — overwhelmed by playing cards / Pokémon cards / business cards):

title:"index cards" AND mediatype:texts

Each IA item bundles a _jp2.zip of all card scans plus ABBYY OCR derivatives, so a single download yields both the images and weak text ground truth.

Citation

Please credit the source institution if you use this dataset:

US Navy Bureau of Medicine and Surgery, History Office. Navy Nurse Corps Index Card collection. Internet Archive: <https://archive.org/details/usnavybumedhistoryoffice>

Sibling archival-card datasets are being progressively released through the BigLAM "Index card datasets" collection.