This document provides a comprehensive description of the structural metadata of the Web Intelligence Hub Online Job Advertisements (WIH-OJA) dataflows. It details the logical organisation of the WIH-OJA and WIH-OJA-NLP databases, including the available databases, tables, record structures, and variable lists used to disseminate OJA microdata for scientific purposes.
The document covers the full set of released table families (main, blended, skills, locations, sources, dates, NLP-derived tables, and codelists), clarifying their scope, relational structure, versioning approach, and intended analytical use. Particular attention is given to tables containing multi-valued attributes (e.g. skills, locations, and sources), as well as to NLP-based outputs such as occupation and skill classifications inferred from job titles and descriptions.
A detailed data dictionary is included, providing definitions, classifications, formats, and methodological notes for all variables. This includes references to international statistical standards and classifications such as ISCO-08, ESCO, NACE Rev.2, ISCED 2011, ISO language and country codes, and NUTS/LAU territorial units.
The document is intended to support statisticians and researchers in understanding, querying, and correctly interpreting WIH-OJA microdata accessed via the WIH DataLab, and to facilitate reproducible analysis, integration with other statistical sources, and informed methodological use of online job advertisement data in labour-market research.
