Semantic Layer and Knowledge Graphs
Graphs have been with us for decades, but a few things have changed since the days of writing algorithms in Matlab:
- Graph stores: Neo4j, TigerGraph, Graphdb, Neptune, RelationalAI – and many more. Native graph databases allow us to build relational models instead of hacking through endless JOINs in traditional databases.
- Graph algorithms: Deep learning is now native to graphs, and even before that, new ways to allow graph alignment and graph embedding helped us to combine, enrich and transfer anonymized knowledge.
- Breakthroughs in Natural Language Processing: Many of us played around with ChatGPT and were amazed by how far NLP had advanced and how rich it spanned across different domains of knowledge. Now imagine “marrying” a relational model with semantic understanding in the knowledge graph.
Given the power of all the above, there has never been a better time to use knowledge graphs to automate and enhance data modeling, management, and discovery.
Leveraging metadata, de facto, the traces of our behavior with data and analytics assets can give us an even bigger win. This win is twofold: the automation of a usage-based data model and the generation of a semantic layer on top of it (Business Terms, Metric Store), as well as the understanding of how users interact with all of that.
The result is Augmented Governance, Assisted Observability, and faster and wiser Analytics.
In illumex, we decided to focus on Semantic Layer Activation to achieve just that.