Automating Job Post Alignment With Esco For Labor Market Insights

The labor market dynamics are followed by policy makers, companies, and the educational sector at large for assessing how well the skills set of workers fits the needs of a given region. The dynamics can be inferred by distinct methods including: (1) interviews and focus groups with stakeholders, and (2) public jobs posts analysis, this later method being non-invasive and thus more suitable for being applied continuously. A key step in analyzing job posts, whether manually or automatically, is to align them with an occupation referential. After such alignment, a clearer view of the labor market’s current state and dynamics is produced. This research presents a novel approach for automated alignment of job posts with the European Skills, Competences, Qualifications and Occupations (ESCO) ontology. The approach requires less data to achieve state-of-the-art performance than previously found in literature and is based on 4 distinct dimensions of analysis. This approach enables a faster adaptation to market changes and trend detection. The algorithm was tested with English jobs posts and was obtained a F1-score of 0.866 for the alignment of job posts with first level ESCO Occupations. Based on the alignment with ESCO we have created dashboards that show distinct dimensions such as the evolution of skills and occupation demands over time, the market share of skills and occupations, the demands for specific geographical regions and/or periods of time.