IAUC Systematic Literature Review: Extraction Protocol, Metadata, and Validation Data

Data and code archive for a systematic review of the urban climate literature, in which large language models extracted structured metadata from the full text of every paper in the corpus. Contents data/extracted_metadata/ — 9,284 canonical per-paper JSON records data/enrichments/ and data/contribution_rationales/ — 9,284 per-paper enrichment and rationale records data/volunteer_review/ — anonymised expert validation responses data/cross_model_validation/ — model comparisons supporting the extraction choice, plus later isolated noncanonical GPT-5.6 audits that do not replace the canonical corpus or feed manuscript results src/, tests/, Makefile, and uv.lock — the analysis pipeline and paper-reproducibility test suite Reproducing the analysis uv sync --extra dev then make paper-all rebuilds the consolidated corpus and manuscript statistics. make ci-test runs the offline tests shipped with the package. Licensing Everything in this deposit — data and code alike — is released under CC-BY-4.0. This version Refreshes the revised-manuscript archive after the final plotting audit and adds portable records from subsequent, explicitly noncanonical extraction comparisons. Canonical Gemini records, corpus membership, and manuscript results are unchanged.