Leiden, 26 May 2026 – Najada Feimi (LISER) showcased SkiLMeeT research at the Joint Research Seminar on Open Job Vacancies Analysis, organised by the SkillsPULSE project.
The online event on 26 May 2026 brought together more than 50 researchers to exchange ideas and address key methodological and data-related challenges shaping the future of job vacancy analysis.
Her presentation, “The Data Processing Pipeline for OJV – Methods and Challenges”, addressed how online job vacancy (OJV) data can be processed to generate reliable insights on skills demand and labour market dynamics.
Drawing on a dataset of more than 100 million job advertisements covering Luxembourg, Belgium, France, and Germany, the presentation highlighted the complexity of working with web-scraped labour market data. Rather than relying directly on existing variables, the SkiLMeeT team developed a comprehensive data processing pipeline to extract key information – including skills, occupations, locations, and firm characteristics—from raw job descriptions.
The process involves multiple stages, including data cleaning, deduplication, information extraction using machine learning models, and validation against manually labelled data. These steps are necessary to address common challenges such as inconsistent information, multilingual content, and differences in data quality across countries.
The presentation emphasised that building reliable indicators from OJV data requires iterative refinement and careful validation at each stage, particularly when aiming to support cross-country analysis and policy-relevant research.
The seminar brought together Horizon-funded projects, researchers, and representatives from EU institutions to discuss how web-based labour market intelligence can contribute to understanding skills needs and supporting evidence-based policy development across Europe.
