JobHop
收藏资源简介:
JobHop数据集由根特大学的研究团队创建,是一个大规模的公共数据集,源自比利时弗拉芒地区公共就业服务机构VDAB提供的匿名简历。该数据集包含超过2.3百万个工作经历,来自391,000多个用户简历,并映射到标准化的ESCO职业代码。使用LLMs技术从非结构化简历数据中提取结构化职业信息,并通过多标签分类模型映射到ESCO代码。数据集支持多样化的应用,如分析劳动力市场流动性、工作稳定性以及职业中断对职业转换的影响。此外,它还支持职业路径预测和其他数据驱动决策过程。数据集通过分析弗拉芒劳动市场的关键特征,如工作分布、职业中断和工作转换,展示了其在推动劳动力市场研究方面的价值。
The JobHop dataset, created by a research team at Ghent University, is a large-scale public dataset sourced from anonymized resumes provided by VDAB, the public employment service of the Flemish region of Belgium. This dataset contains over 2.3 million work experiences from more than 391,000 user resumes, and has been mapped to standardized ESCO occupational codes. Structured occupational information is extracted from unstructured resume data using large language models (LLMs), and mapped to ESCO codes via a multi-label classification model. The dataset supports a diverse range of applications, including analyses of labor market mobility, job stability, and the impact of career interruptions on job transitions. In addition, it also supports career path prediction and other data-driven decision-making processes. By analyzing key features of the Flemish labor market such as job distribution, career interruptions and job transitions, the dataset demonstrates its value in advancing labor market research.
数据集概述:JobHop
基本信息
- 许可证: CC BY 4.0
- 语言: 英语 (en)
- 标签: labor_market, ESCO, Job_transition
- 数据规模: 1M < n < 10M
数据集描述
- 数据来源: 来自比利时弗拉芒公共就业服务机构VDAB提供的超过40万份匿名简历。
- 创建方法:
- 使用小型语言模型(LLM)从简历中提取信息。
- 使用Nobl.ai提供的专有工具将每段工作经历映射到对应的ESCO代码。
- 包含字段:
- person_id: 标识该行数据所属的简历
- esco_code: 匹配的ESCO代码
- start_date, end_date: 工作经历的起止日期
- university_degree: 个人是否拥有大学学位的标志
作者与资助信息
- 作者: Iman Johary, Raphael Romero, Alexandru-Cristian Mara, Tijl De Bie
- 资助方:
- 根特大学BOF (BOF20/IBF/117)
- 弗拉芒政府 (AI Research Program)
- FWO (11J2322N, G0F9816N, 3G042220, G073924N)
- ERC grant (VIGILIA, 101142229)
引用与许可
- 引用信息: 即将添加
- 许可证: CC BY 4.0
联系方式
- 通讯作者: iman.johary@ugent.be




