Publish-Review-Curate Platform Dataset - harvested source records (Crossref works and peer reviews, Sciety docmaps)

This archive contains the harvested source records behind one extraction run (run_20261102_01) of the Publish–Review–Curate (PRC) Platform Dataset. The dataset is register-based: it gives annual counts of unique research works entering publish-first and open-review workflows, and of works receiving public review or curation after initial publication, compiled from the Crossref REST API and Sciety docmaps. Its rows are observation series rather than organisations and are classified on two axes: initiation class and PRC configuration. Every count is published with its eligible year range and an explicit data status. The repository tracks the configuration, documentation, data-quality log, page manifests, and published tables and figures. Harvested third-party records are not tracked there; under decision D-056, they are deposited here instead. The archive holds data/processed/records_*.csv (normalised source records for each of the 41 register rows with a completed extraction — one file per row; the 3 register-only rows without a machine-readable source have none, and the PeerRef file carries the header only because its Sciety group is absent from the docmaps index, D-058); works_included.csv (work-level inclusion, exclusion reasons, counting dates and the Gates Open Research–VeriXiv correspondence basis); normalisation_summary.json; analysis_manifest.json; the interim extraction files in data/interim/; and a copy of data/manifests/run_20261102_01/, which records the SHA-256 digest and terminal state of every accepted page and walk, the Sciety index snapshot used to verify source coverage, and the OpenAlex diagnostic walks of D-066 (the OpenAlex responses themselves are not redistributed). SHA256SUMS.txt lists the SHA-256 digest of each of the 631 files. The SHA-256 digest of the archive itself is recorded in the repository at reports/release_archive_run_20261102_01.sha256. This run supersedes run_20261027_01 (10.5281/zenodo.23812673). Between the two runs the pipeline gained the posting-record gate for eLife Reviewed Preprints and Access Microbiology (D-060; the Access Microbiology coverage start is resolved from the run as 2022, D-017, re-derived by every complete run, not written to the register), the Gates Open Research–VeriXiv pairing from Crossref is-preprint-of/has-preprint relations (D-041), Crossref peer-review records as the primary source for The Unjournal, MetaROR and Review Commons (D-063–D-065), group cells published as lower bounds rather than blanks (D-061), windowed Crossref walks for scopes with cursor-step loss (D-057), and the OpenAlex volume check and pair signal as data-quality rows (D-066). Crossref records and Sciety docmaps were retrieved on 2026-11-02 (walks completed between 08:38 and 12:08 UTC; 424 scope entries in the run manifest, of which 55 are prefix, ISSN or group scopes and 369 are the monthly or quarterly window sub-scopes of D-057 for three peer-review scopes). The records are redistributed as retrieved and reflect the state of the sources on that date. The archive corresponds to repository commit d0b4806 (later commits touch documentation only) and dataset version 1.4.0. Counts, methods, limitations and the decision log are documented in the repository; this deposit exists so that every published number can be traced back to the retrieved records. Third-party content. Bibliographic metadata retrieved from the Crossref REST API (DOIs, record types, dates, relations, titles, container titles) is factual and redistributed as retrieved; Crossref makes member-deposited metadata available for reuse without restriction (terms recorded in config/sources.yaml). Docmap metadata retrieved from Sciety (DOIs, dates, evaluating groups, evaluation URLs) is redistributed with attribution to Sciety (eLife Sciences Publications, Ltd). No abstracts or evaluation texts are included. The compilation, normalised fields, inclusion decisions and page manifests are the author's work under CC BY 4.0. The analytical code was written with the assistance of Perplexity. All design, classification and data-handling decisions were taken by the researcher, who reviewed the code and its outputs and bears final responsibility for the integrity of the data.