Metadata based extraction of data insights

Hi All,

I am trying to implement a system for which I need some documents/references.

Background -

  • We have data lake which has organizational data.
  • We need to build a system which allows us to -
    • Query the data,
    • Extract insights from the data,
  • We do not want to replicate the data anywhere, because it would be redundant practice.
  • Reference diagram is shared below.

I have checked few articles but could not find one which can help related to the reference diagram provided, usually while using knowledge graph we put data inside it.

What I need,

  • If anyone has any links, references, videos, presentations, documents, any other kind of source which can help us build such architecture.
  • Help us with the feasibility check over such system.

Reference Diagram -

Thank you.

If you have unstructured data (“documents” like doc / pdf / xls ) - you might need to use something like Docling to extract the information to build your KG.

Suppose you have JSON or other types of structured data. In that case, you might need to categorise it and create “pipes” that add/generate nodes from the information (as it is ‘precise’, you don’t want the probabilistic element of interpretation, but you can if you don’t mind losing some precision).