How to move towards Knowledge Graphs?

[davidgasquez]

Using KGs could help users connect datasets (detecting and merging on Common Columns or using something like Ver). With good enough core datasets, it could be used as an LLM backend.

Is it possible to build a community curated Knowledge Graph of Frictionless DataPackages?

• Normalize tabular data down to clean, well-formed CSVs or data that is more graph-like.
• Link datasets with RDF / JSON-LD. JSON-LD context for Data Package and Tabular Data Package frictionlessdata/datapackage#218

There might be many more ways to build semantic datasets too (like using LLMs)!

[davidgasquez]

Some related and potentitally insteresting resources:

• Plow.
• Some Frictionless discussions:
  • Reconciling with package.jsonld frictionlessdata/datapackage#110
  • JSON-LD context for Data Package and Tabular Data Package frictionlessdata/datapackage#218
• Atomic Data.
  • They considered IPFS.