Data and Analysis Materials for "Generalizability and Evidence Maturity of Machine Learning Models for Urban Flood Susceptibility Mapping: A Systematic Review and Multilevel Meta-Analysis"

This repository contains the final data, coding documentation, and analysis materials supporting the systematic review and multilevel meta-analysis titled “Generalizability and Evidence Maturity of Machine Learning Models for Urban Flood Susceptibility Mapping: A Systematic Review and Multilevel Meta-Analysis.” The locked evidence base comprises 176 included studies and 1,079 model-evaluation records. The AUC analytical dataset contains 607 extracted AUC estimates contributed by 115 studies. The repository includes: (1) the final study-level dataset and model-evaluation records; (2) the AUC analytical dataset; (3) a comprehensive data dictionary; (4) the coding manual defining eligibility, urban-basis classification, validation-design taxonomy, AUC evidence tiers, reproducibility indicators, risk-of-bias assessment, and the Generalizability and Evidence Maturity framework; and (5) the R analysis script supporting the reported statistical analyses. The final database search endpoint was July 10, 2026. Missing or unreported information is coded according to the definitions provided in the data dictionary and coding manual. The repository is intended to support transparency, verification, and reuse of the review’s coded evidence base and analytical procedures.