EnvCoRe-SW: quality-controlled environmental monitoring measurements and a de-identified inventory of compliance reports from Southwest China (2018–2025)

EnvCoRe-SW v6 is a de-identified research dataset derived from environmental compliance monitoring reports collected in Guizhou Province, Southwest China, during 2018–2025. It contains 20,023 structured environmental measurement records linked to 8,259 public report-inventory records. The release includes the frozen Stage 6D-R2 measurement and inventory outputs, controlled vocabularies, a data dictionary, a pollutant dictionary, known-issue records, a public file manifest, checksums, and aggregate Final300, Probability50, and schema-scope validation results. The public distribution contains 19 files. The dataset supports environmental data curation, structured measurement reuse, parameter harmonization, provenance-aware quality assessment, validation research, and environmental informatics. It does not include original source reports, complete report text, organization or facility names, project names, report numbers, source paths, private crosswalks, internal identifiers, source evidence, or reviewer identities. Final300 evaluates record-level fields, Probability50 evaluates whole-report completeness within the frozen parameter schema, and the schema-scope analysis quantifies source-supported environmental results that fall inside or outside that schema. Assessable accuracy estimates should be interpreted together with NOT_ASSESSABLE counts, partial-identification intervals, sampling designs, and documented review limitations. No dataset-wide 100% accuracy claim is made. This dataset is intended as a research data resource rather than an independent regulatory compliance determination. Users should consult the data dictionary, controlled vocabularies, known-issue table, scope and limitations, usage notes, and validation summaries before analysis. The dataset should not be used to attempt re-identification of source organizations, facilities, projects, or reports.