Coded dataset and analysis scripts for: Diffusion Models for Image Restoration and Medical Imaging — A PRISMA-Guided Systematic Review, Bibliometric Analysis, and Evidence Mapping of 874 Studies

Complete evidence base for a PRISMA-guided systematic review of diffusion models for image restoration and medical imaging. Contains 961 de-duplicated bibliographic records (1998–2025), of which 874 core studies are coded against a fixed 17-field schema (method family, task, domain, medical modality, datasets, metrics, supervision regime, sampling budget, stated limitation, and more), in both XLSX and JSON form; the completed PRISMA 2020 checklist; the multi-label thematic coding of 235 stated limitations; and all Python scripts that reproduce every figure, table and statistical result in the manuscript (Wilson confidence intervals, trend-model comparison with change-point bootstrap, code-release trend tests, consistency checks), plus a unified-protocol GPU benchmark harness. Full-text PDFs are not redistributed; each record carries identifiers sufficient to retrieve the source.