Research on the Construction and Dynamic Evolution Analysis of Job and Competency Graphs Driven by Multi-Source Heterogeneous Data

DOI: https://doi.org/10.62517/jike.202604315

Author(s)

Huaruo Mao*, Zexiang Zhang, Yifei Zhai, Zihao Wang, Yuexin Wang, Ruian Yan

Affiliation(s)

School of Artificial Intelligence and Big Data, Henan University of Technology, Zhengzhou, Henan, China *Corresponding Author

Abstract

With the rapid development of emerging information technologies such as artificial intelligence and big data, industrial structures and talent demands are continuously changing. Traditional talent analysis methods based on fixed job classification and keyword retrieval are difficult to meet the demands of the employment market. This paper designs and implements an AI job intelligent analysis platform driven by multi-source heterogeneous data. The system focuses on four core aspects: job data collection, job knowledge construction, job dynamic analysis, and intelligent person-job matching. Through Scrapy and Playwright, the system realizes automatic multi-source data collection, and combines natural language processing, large language models, and knowledge graph technologies to achieve job information analysis, skill entity extraction, and dynamic updates of job competencies. The system uses MySQL and Elasticsearch for structured data storage and full-text retrieval, utilizes Neo4j to construct job-skill-industry knowledge associations, and implements RAG-based job definition generation through LangChain and ChromaDB. The platform realizes functions including multi-source job data management, new job discovery and definition, job competency dynamic evolution analysis, job panorama graph visualization, intelligent resume parsing, and person-job matching recommendation. The system adopts a front-end and back-end separated architecture, with the front-end built based on Vue3 to construct visual interactive pages and the back-end based on FastAPI to provide interface services. The test results show that the platform can effectively support job data collection, job knowledge management, job competency analysis, and talent matching requirements, providing intelligent decision support for enterprise recruitment and individual career development.

Keywords

Multi-Source Heterogeneous Data; Job Competency Graph; Large Language Model; Knowledge Graph; Job Dynamic Evolution; Person-Job Matching

References

[1]ZHAO Peng. Deeply Exploring the Value of Archival Information and Activating Precise Person-Job Matching. Human Resources, 2026, (13): 74-75. [2]SHAO Xintian, CHEN Junli, ZHANG Hanjv. Design of a Server Cluster Monitoring and Evaluation System Based on Elasticsearch. Computer Technology and Development, 2025, 35(9): 77-84. DOI:10.20165/j.cnki.ISSN1673-629X.2025.0120. [3]RANGE J, SCHEMBERA B, GÖDDEKE D. Making Mathematical Knowledge Explainable, Accessible and Interoperable Through Large Language Model Integration. arXiv preprint, 2026. arXiv:2607.24512. [4]ZHU Yan, WANG Hailin. Design of Personalized Recommendation Algorithm for Vocational Education Resources Based on Dynamic Knowledge Graphs and Reinforcement Learning . Journal of Jilin University (Information Science Edition), 1-9 [2026-07-28]. [5]FENG Li. Research on Chinese Resume Information Entity Recognition Based on Deep Learning. Wuhan Polytechnic University, 2024. DOI:10.27776/d.cnki.gwhgy.2024.000493. [6]CHEN Cui. Research on the Application of Artificial Intelligence in Enhancing Enterprise Recruitment Matching Accuracy. Market Weekly, 2026, 39(19): 187-190. [7]YU X, XU R Z, XUE C Y, et al. ConFit v3: Improving Resume-Job Matching with LLM-based Re-Ranking. arXiv preprint, 2026. arXiv:2605.09760. [8]JOUANNEAU W, JOUFFROY E, PALYART M. An Efficient Long-Context Ranking Architecture With Calibrated LLM Distillation: Application to Person–Job Fit. arXiv preprint, 2026. arXiv:2601.10321. [9]QIAN Lei, LONG Xinyue. Research on Personalized Learning Path Design and Teaching Adaptability Enabled by AIGC. China Computer & Communication, 2026, 38(16): 230-232. [10]HUANG Liming. Research on Optimization of Multi-source Heterogeneous Data Fusion Storage Architecture. Network Security Technology & Application, 2026, (6): 85-88.