An open-data team wants methodology sections from statistical report PDFs turned into structured outputs for cataloging. What should analyzers produce?

An open-data team cataloging statistical report PDFs should have analyzers produce structured outputs that capture methodology sections so those sections can be reused as catalog metadata and evidence. Methodology defines sampling, definitions, and limitations that users of municipal open data need; structure makes those passages queryable and assignable to dataset records. Structured outputs that summarize methodology sections from the statistical report PDFs condense sampling, definitions, and limitations into paraphrases instead of capturing the stated text, losing the complete evidence needed for catalog reuse. Structured outputs that capture methodology footnotes from the statistical report PDFs pull only ancillary notes instead of the full methodology sections, leaving sampling, definitions, and limitations incomplete for catalog records. Vector embeddings that capture methodology sections from the statistical report PDFs produce numeric vectors for similarity search instead of structured outputs, leaving methodology text unmapped to queryable catalog fields. Exam caveat: open-data methodology reuse from statistical PDFs requires structured analyzer outputs for those sections—not summarizing methodology sections from the statistical report PDFs, capturing methodology footnotes from the statistical report PDFs, or vector embeddings that capture methodology sections from the statistical report PDFs. Operational check: configure section-aware structured extraction targeting methodology headings; map outputs to catalog fields; review a sample of datasets for complete methodology text; reject configurations that summarize methodology sections from the statistical report PDFs for this cataloging workflow before portal publish.