A specialized course focusing on integrating and managing data using semantic web technologies at an enterprise scale. It is designed for data architects and engineers who face the challenge of combining data from disparate sources and need a more flexible, semantic solution. The course covers how knowledge graphs act as a unifying data layer that provides flexible data integration across silos. Participants will learn to convert relational/legacy data to RDF, align schemas via ontologies, and use semantic standards to break down data silos. This training empowers professionals to build an enterprise knowledge graph that ensures consistent, machine-understandable data across the organization. (Attendees should have intermediate knowledge of semantic tech or databases.)
Curriculum:
Module 1: Principles of Semantic Data Integration – Understanding enterprise data silos and integration challenges; how semantic frameworks (RDF/OWL) provide a common data model to unify heterogeneous data and enable interoperability.
Module 2: Ontologies for Data Integration – Using ontologies and vocabularies as a mediated schema for integration; mapping disparate data schemas to a common ontology, and strategies for ontology alignment and data federation.
Module 3: Data Conversion and Mapping – Techniques for converting various data formats to RDF: using ETL tools and standards like R2RML (Relational to RDF Mapping) to transform databases into knowledge graphs, and mapping CSV/JSON data sources to RDF.
Module 4: Entity Reconciliation and Linking – Methods for linking records referring to the same real-world entity across datasets using semantic approaches (URI linking, sameAs relationships), leveraging reference knowledge bases to reconcile entities, and maintaining data consistency.
Module 5: Semantic ETL and Tools – Overview of pipelines and tools for semantic data integration (such as Karma, Apache Jena’s riot tools, or enterprise platforms like Stardog), and how to build repeatable ETL processes that populate and update a knowledge graph.
Module 6: Enterprise Knowledge Graph Management – Best practices for maintaining an enterprise knowledge graph: data governance, handling updates and versioning of data and ontologies, ensuring data quality with validation (SHACL/ShEx), and performance considerations for querying integrated data at scale.
Module 7: Integration Project – A practical project where learners take multiple example datasets (e.g., from different departments or sources) and integrate them into a unified knowledge graph, demonstrating the end-to-end process of semantic data integration and answering cross-source queries that highlight the power of integration.
