Abstract: A semantic data processing system for processing an unstructured data file based on semantics of textual content of the unstructured data file comprises a processing arrangement configured to: receive, structured reference data, via reference data ingester; receive unstructured data files via an unstructured data file preprocessor: generate, via knowledge graph generator, knowledge graph from the structured reference data using structure of the structured reference data, the knowledge graph specifying a plurality of different entities having relationships with each other and relationship data linking the different entities; generate, via entity-centric search query generator, entity-centric search query by using the knowledge graph, the entity-centric search query specifying plurality of entity-related textual terms and parameters regarding the relationships between the entity-related textual terms; and execute, via contextual matching engine, the entity-centric search query on the received unstructured data