Why indexing is needed
The Phase 1 source is the 1909 Library of Congress Cherokee Nation atlas. The maps can show names, statuses, and allotment numbers, but those details are in map images. To search by family name or allotment number, the map text has to be transcribed, checked, and tied back to the original map page.
Safe transcription workflow
- Choose a map page. Start with township and range, such as T21N R12E.
- Open the original source. Review the map at the Library of Congress.
- Transcribe only what appears on the historical public record. Do not add private family stories or living-person information.
- Record uncertainty. Use question marks, notes, and confidence levels when handwriting or map text is unclear.
- Review before publishing. A row should not become a public searchable family result until it has source links and review status.
Fields to capture
Roll/enrollment numbers should be added only when supported by Dawes/NARA, allotment jacket, or other source-linked records. The LOC map layer alone should not pretend to be a complete roll-number database.
Map Indexing Agent starter kit
Version 0.8 adds a local OCR helper that can download one public LOC map image, split it into tiles, run Tesseract OCR, and create candidate rows for human review. It does not publish records automatically.
Run the agent locally first. Public submissions and automatic publication remain closed. Every OCR candidate must be reviewed against the original source map before it becomes a searchable public record.
No public upload form yet
Public submissions remain closed until privacy, consent, review, and removal processes are finalized. For now, this page explains the data plan so the project can grow carefully.