# Explore large-scale data acquisition

Read https://textdata.info/toolkit/large-scale-data-acquisition/ and its working files at https://github.com/joyzhzh/textdata/tree/main/appendix/08-data-acquisition.

Start with README.md, tool/WORKFLOW.md and docs/PROCEDURE.md. Use example/run_example.py and the records under example/output/ to explain the 7–8 December 2020 example: define the document population, retain gaps, prepare reading copies, package files and compare fingerprints. Show how the webpage’s year/month explorer and data/MANIFEST_INDEX.tsv could help define a sampling frame for my question.

If local files and Python are available, help me run the documented offline example and interpret its actual output. Otherwise explain the supplied example without claiming you ran it. Explain that the historical comparison found 191 identifiers, 169 raw matches and 22 unresolved differences; the supplied email-redacted copies produce a separate comparison. Distinguish byte integrity from content validity.

Keep completed work separate from proposed integration and additional-source adapters. Note that full historical text packs are not supplied, full historical acquisition cannot be replayed, images may be missing, and the demonstration covers one provider and two publication days. Cite the files you use and suggest a concrete next step for my question.

My research question or task: [describe it here]