Virtual Knowledge Graphs (VKGs) have recently gained attention due to their flexible data model, which reduces the effort needed for integration across different, possibly heterogeneous, data sources.

In this webinar, Alessandro Mosca presents how to access scientific data stored in a (relational) database through the Virtual Knowledge Graph (VKG) approach. In such an approach, the data are exposed as a Knowledge Graph and enriched with semantic information coming from a domain ontology. The VKG approach is introduced by means of real-world scenarios, involving both scientific and business data and using the open-source VKG system Ontop.

Since legacy data are exposed as a Knowledge Graph, users can access the data by means of a more convenient vocabulary provided by the domain ontology, benefit from automated reasoning capabilities, and do not need to focus on how the data are actually stored. By relying on existing federation tools, Alessandro shows that the VKG approach can also be used to integrate multiple, heterogeneous, and possibly semi-structured and unstructured data sources.

Speaker

Alessandro Mosca
Assistant Professor at the Free University of Bozen-Bolzano

Questions

00:42:55 How to update the ontology in the case of VKG, if it does not cover all data sources ? – Sarra Ben Abbes
0:49:56 SPARQL looks rather similar to Cypher. Would a drag-and-drop GUI interface to generating queries be useful for non-experts? (Neo4j has an app that can be used that way, to some extent) – Kimberly Cordwint Martin
00:53:01 Most of the data in agriculture are in Excel file. Would the mapping approach work with Excel/CSV files? – Marie-Angélique Laporte
00:55:30 What is the advantage of using virtual KG versus materialized? I can see the advantage, as you said, when data are updated often, but could you tell us more in terms of performance? – Marie-Angélique Laporte
00:59:30 Is it good to combine the material and virtual approach? – Lynda Temal
01:00:10 What you can advise to handle time series data (like sensor data) – Lynda Temal
01:01:57 You spoke a lot about the RDBMS scenario and the associated schemas. What about No-SQL approaches that tend to be ‘schema-lite’ or even schema free? Crawford Revie
01:03:55 – For the materialized KG, do you have a specific triplestore to recommend for big semantic data? – Sarra Ben Abbes