Construction Safety Dataset (CSDataset)
收藏资源简介:
CSDataset 是一个由亚利桑那州立大学创建的大型多层级建筑安全数据集,包含来自职业安全与健康管理局 (OSHA) 的 2013 年至 2022 年间超过 50,000 条事故记录、100,000 条检查和关联违规数据。该数据集整合了结构化属性(如事故类型、伤害严重程度、工作职位、天气状况和地理数据)和非结构化叙述字段,支持机器学习和大型语言模型的各种方法,可用于事故严重程度预测、风险分析等任务。
CSDataset is a large multi-level construction safety dataset created by Arizona State University. It contains over 50,000 incident records and 100,000 inspection and associated violation records sourced from the U.S. Occupational Safety and Health Administration (OSHA) between 2013 and 2022. This dataset integrates structured attributes such as incident type, injury severity, job role, weather conditions, and geospatial data, alongside unstructured narrative fields. It supports a diverse range of machine learning and large language model (LLM) methodologies, and can be utilized for tasks including incident severity prediction, risk analysis, and more.

- 1Building Safer Sites: A Large-Scale Multi-Level Dataset for Construction Safety Research亚利桑那州立大学 · 2025年



