Dataset for "A Probabilistic Framework for Spatio-Temporal Social Vulnerability Assessment to Flood Hazards in Canada using Bayesian Machine Learning"

This dataset provides the foundational geospatial and socioeconomic data used in the manuscript "A probabilistic framework for spatio-temporal social vulnerability assessment to flood hazards in Canada using Bayesian machine learning" (Chakraborty et al., 2025), accepted and published in the Springer Nature "Journal of Geovisualization and Spatial Analysis" (DOI: 10.1007/s41651-025-00247-y). It offers a comprehensive suite of indicators designed to quantify social vulnerability through advanced computational methods, including Bayesian machine learning (ML), mutual information regression (MIR), and Monte Carlo simulations. This dataset ensures maximum transferability of the research and allows for the precise reproducibility of empirical results. Users are responsible for verifying the accuracy and applicability of the data for their specific needs. Redistribution or commercial use of the data is subject to licensing restrictions from original data providers. Cite appropriately when using this dataset in other publications or presentations. Key Components: Geospatial Layers: Shapefiles (.shp) containing administrative boundaries and spatial attributes. Socioeconomic & Demographic Indicators: Processed variables from Statistics Canada, including income, housing, and education metrics. Racial/Ethnic Variables: Specific data layers used to assess distributive environmental justice and disproportionate risk. Documentation: A detailed data dictionary defining all variables and attributes. Research Application This dataset is curated for researchers and practitioners working in: Dynamic Social Vulnerability Analysis: Monitoring how vulnerability shifts over time and space. Probabilistic Risk Assessment: Incorporating uncertainty into flood risk modeling. Environmental Justice: Evaluating the distributive impacts of natural hazards on marginalized populations. Emergency Management: Supporting evidence-based, proactive disaster risk reduction strategies in Canada. Technical Details Software Requirements: Data processing was conducted using ArcGIS Pro and Python. Source Attribution: Census data are obtained from open-source Statistics Canada products; Canada flood hazard data are copyrighted by JBA Risk Management; and residential address point data are copyrighted by the DMTI CanMap Suite. Contact Information For inquiries regarding the methodology or data application, please contact the corresponding author: Dr. Liton Chakraborty University of Waterloo: [e-mail], or York University: [e-mail]本数据集为发表于施普林格自然（Springer Nature）旗下《地理可视化与空间分析期刊》（Journal of Geovisualization and Spatial Analysis，DOI: 10.1007/s41651-025-00247-y）的论文《基于贝叶斯机器学习的加拿大洪涝灾害时空社会脆弱性评估概率框架》（Chakraborty等，2025）所使用的基础地理空间与社会经济数据。本数据集提供一套综合指标体系，可通过先进计算方法量化社会脆弱性，所用方法包括贝叶斯机器学习（Bayesian machine learning, ML）、互信息回归（mutual information regression, MIR）与蒙特卡洛模拟（Monte Carlo simulations）。本数据集可最大程度保障研究成果的可迁移性，并支持实证结果的精准复现。使用者需自行验证该数据在其特定场景下的准确性与适用性。数据的再分发或商业使用需遵守原始数据提供者的授权限制。若在其他出版物或演示中使用本数据集，请进行规范引用。 核心组成部分： 地理空间图层：包含行政边界与空间属性的形状文件（.shp）。 社会经济与人口统计指标：取自加拿大统计局（Statistics Canada）的经处理变量，涵盖收入、住房与教育相关指标。 种族/族裔变量：用于评估分配式环境正义与不均等风险的专属数据图层。 文档说明：包含所有变量与属性定义的详细数据字典。 研究应用场景： 本数据集面向从事以下领域研究与实践的人员： 动态社会脆弱性分析：监测脆弱性随时间与空间的动态变化。 概率风险评估：将不确定性纳入洪涝风险建模流程。 环境正义：评估自然灾害对弱势人群的分配性影响。 应急管理：为加拿大基于证据的主动式灾害风险减缓策略提供支撑。 技术细节： 软件需求：本数据集的处理采用ArcGIS Pro与Python完成。 来源标注：普查数据取自加拿大统计局（Statistics Canada）的开源产品；加拿大洪涝灾害数据版权归JBA风险管理（JBA Risk Management）所有；住宅地址点数据版权归DMTI CanMap套件（DMTI CanMap Suite）所有。 联系方式： 若需咨询研究方法或数据应用相关问题，请联系通讯作者：滑铁卢大学利顿·查克拉博蒂博士，邮箱：[e-mail]；或约克大学，邮箱：[e-mail]。