The article presents a formally grounded approach to the development of ontology-driven systems for multidimensional analytics. In this approach, a formal domain ontology is treated as the primary and invariant knowledge model that provides a sufficient basis for the inductive construction of analytical structures. Un-like traditional solutions, where ontologies are typically applied only as auxiliary semantic or integration layers, the proposed framework derives analytical dimensions, hierarchies, indicators, and permissible aggregation operations directly from the logical organization of the ontological model. The study formalizes several key processes required for building such systems. These include the structuring and semantic alignment of heterogeneous data sources, the construction of a formal domain ontology, and the subsequent use of this ontology to support multicriteria evaluation and the generation of multidimensional analytical representations. Within this framework, a mathematical definition of an ontology-analytic mapping operator is introduced. This operator ensures a rigorous transformation from the ontological model to a multidimensional analytical structure while preserving type information, semantic constraints, and interpretive properties. The results demonstrate that analytical facts can be induced from ontological elements, which enables the correct handling of context-dependent and partially defined dimensions. In addition, the article proposes a conceptual architecture for an ontology-driven analytical system organized into functional and conceptual layers. This layered architecture supports methodological consistency, reproducibility of analytical procedures, and scalability of the proposed approach. The presented formalization is general in nature and can be applied to the development of intelligent analytical systems in various domains. In particular, it is well suited for systems designed to analyze and evaluate educational and intellectual achievements, where consistency of semantics, criteria, and analytical procedures is critically important. Overall, the obtained results provide a formal foundation for constructing coherent, semantically consistent, and explainable analytical systems across diverse application areas.The article presents a formally grounded approach to the development of ontology-driven systems for multidimensional analytics. In this approach, a formal domain ontology is treated as the primary and invariant knowledge model that provides a sufficient basis for the inductive construction of analytical structures. Un-like traditional solutions, where ontologies are typically applied only as auxiliary semantic or integration layers, the proposed framework derives analytical dimensions, hierarchies, indicators, and permissible aggregation operations directly from the logical organization of the ontological model. The study formalizes several key processes required for building such systems. These include the structuring and semantic alignment of heterogeneous data sources, the construction of a formal domain ontology, and the subsequent use of this ontology to support multicriteria evaluation and the generation of multidimensional analytical representations. Within this framework, a mathematical definition of an ontology-analytic mapping operator is introduced. This operator ensures a rigorous transformation from the ontological model to a multidimensional analytical structure while preserving type information, semantic constraints, and interpretive properties. The results demonstrate that analytical facts can be induced from ontological elements, which enables the correct handling of context-dependent and partially defined dimensions. In addition, the article proposes a conceptual architecture for an ontology-driven analytical system organized into functional and conceptual layers. This layered architecture supports methodological consistency, reproducibility of analytical procedures, and scalability of the proposed approach. The presented formalization is general in nature and can be applied to the development of intelligent analytical systems in various domains. In particular, it is well suited for systems designed to analyze and evaluate educational and intellectual achievements, where consistency of semantics, criteria, and analytical procedures is critically important. Overall, the obtained results provide a formal foundation for constructing coherent, semantically consistent, and explainable analytical systems across diverse application areas.