METER-AI: Full-Text Literature Corpus for Meteorological Data Representation in Energy AI

The original literature discovery process identified 1,654 candidate publications through searches involving OpenAlex, Scopus, and Semantic Scholar. Following screening and full-text acquisition, a registered collection of 400 research documents was assembled. Dataset organisation The archived PDF collection contains three directories: included/ — 352 documents included in the literature review. excluded/ — 48 documents excluded during eligibility screening. shortlisted/ — 60 documents selected for the detailed scientific synthesis. The 60 shortlisted documents are a subset of the 352 included documents. Consequently, the archive contains 460 PDF copies representing 400 distinct registered records. Reproducibility The accompanying research code and metadata records document the original paper identifiers, eligibility decisions, passage provenance, evidence extraction procedures, and selection of 60 core studies. The archive preserves the document organisation used in the METER-AI review. Its inclusion does not constitute an endorsement of the findings or methods of individual publications. Rights and attribution The PDF collection consists of publications authored by third parties. Copyright and reuse permissions remain subject to each document's original rights holder and applicable licence. The dataset compiler does not claim authorship or copyright ownership of the collected publications.