Reproducible Quantitative Typology of Canton Export Fans — Supporting Data, Scripts, and Audit Outputs (v1.8)

This deposit provides the complete derived dataset, analysis scripts, reproducibility validation report, and audit outputs that accompany the PLOS ONE methodology paper "A reproducible framework for quantitative typology of traditional craft objects: morphometric grouping of measurable images of Canton export fans" (Zhang, Wei & Zheng, 2026). Contents: derived measurement tables for the 186-image collection, 137 unique images, 96 measurable images, and 77 analysis specimens; major-source-balanced resampling output (1000 iterations); k = 4 partition descriptives and BCa bootstrap 95% confidence intervals on group means (10,000 within-group resamples, seed 20260515); log-ratio PCA supplementary check (PCA on standardized log W and log L without W/L as a third variable; ARI 0.7726 vs. the main k = 3 partition); the independent validation report confirming byte-level CSV and pixel-level figure equivalence under independent re-execution; Python scripts that regenerate every figure and statistical result; representative audit overlay images. Raw collection photographs are NOT redistributed because licensing varies by holding institution; this deposit provides collection identifiers, derived measurements, and complete code so external readers can re-execute every numerical result and audit each measurement step. Reproducibility: Python 3.13.7; requirements.txt pinned to exact library versions; random seed 20260515 for all stochastic procedures. Funding: National Social Science Fund of China (Award 2019BKS140); Zhengzhou University Local "101" Curriculum-Building Project. License note: Derived data, documentation, and figures are released under CC-BY 4.0. Python scripts in the scripts/ folder are dual-licensed under MIT (see LICENSE-CODE in the deposit).