Notes on the MEDem ATLAS Prototype

This short note introduces the MEDem ATLAS prototype: an online application designed to demonstrate the potential of concept-driven dataset annotation for facilitating comparative social science research. The system implements a structured data model linking an ontology of theoretical concepts to tabular datasets containing variables and metadata. Within this framework, concepts for object properties—such as attributes of persons or political parties—are defined and organized hierarchically into dimensions, subdimensions, and indicator concepts, each supported by literature-based definitions. These conceptual elements are systematically connected to empirical operationalizations through a variables database that records how indicators are measured across datasets. By linking conceptual knowledge to dataset metadata, the tool enables interactive exploration of conceptual coverage across datasets and facilitates the identification and comparison of available operationalizations. The system also includes a Research Builder tool that allows users to select concepts relevant to a research question and automatically identify datasets containing corresponding indicators, while providing detailed information on measurement coverage. The prototype further demonstrates the possibility for cross-dataset linkage and generates automated scripts for preliminary analysis in common data science languages. Overall, MEDem ATLAS illustrates how ontology-based annotation can enhance transparency, comparability, and efficiency in secondary data analysis.