A harmonized fast-fashion garment-variant dataset for textile circularity and sustainability assessment

This record contains the complete archived release of a harmonized fast-fashion garment-variant dataset and preprocessing workflow derived from publicly accessible online product pages of H&M and Uniqlo in the United Kingdom and Australia. The final public dataset file, 6_JSONL_component_normalized_public.jsonl, contains 47,522 colour-specific garment variants. Each record includes harmonized product metadata, source URL and timestamp information, colour information, retailer-disclosed material composition, normalized material text, harmonized garment categories, and structured component-level material-composition information. The public dataset was derived from the component-normalized processing output by removing the non-essential consumer-engagement fields rating and reviewCount from all records. The release also includes six preprocessing scripts, five CSV documentation tables, six TXT processing summaries, and a public-curation summary documenting the transformation from retailer product-page records to the final public component-normalized dataset. The associated GitHub repository contains the maintained documentation, preprocessing scripts, mapping tables, processing summaries, and workflow description. The dataset supports garment-level analysis of textile circularity, automated textile sorting, textile pre-processing requirements, garment construction barriers, and fibre-to-fibre recycling potential. It is associated with the co-submitted Resources, Conservation and Recycling manuscript “Garment construction creates overlooked barriers to textile sorting and fibre-to-fibre recycling in fast fashion.” The public release provides a curated research dataset rather than a redistribution of retailer webpages. It does not redistribute product images, model images, screenshots, raw HTML, full webpage captures, direct image URLs, review text, product ratings, review counts, unnecessary commercial webpage material, or retailer-specific acquisition scripts. Brand names are used only to identify the source websites from which publicly accessible product information was obtained. H&M and Uniqlo are trademarks of their respective owners; the authors are not affiliated with, endorsed by, or sponsored by H&M, H&M Group, Uniqlo, or Fast Retailing.