Create Knowledge Graph
This document is step by step guide to build a Knowledge Graph
Prerequisites
A GraphDB connector must be configured.
A VectorDB (MongoDB, pg-vector, Pinecone, or OpenSearch) must be configured.
A SearchEngine connector must be configured.
An Embedding Model must be configured.
After logging in to Dataworkz, click Knowledge Graph from the top menu.
Click the Add button and select Create Knowledge Graph. Follow the three steps below to build a Knowledge Graph.
Step 1 — Basic Configuration
Goal: tell Dataworkz what data to read and which entities to extract.
Enter a name for the knowledge graph.
Select the graph database instance where the Knowledge Graph will be written.
Select a graph template from the list. A template contains prompts from the Prompt Library.
Select the LLM to use for entity extraction.
Select the source dataset.
Select the target where extracted data will be written:
Select the Workspace.
Select the Collection.
Select an existing directory or create a new one.
Click Next.
Step 2 — Advanced Options
Goal: configure where embeddings and lexical indexes are stored and which models are used.
Vector Database & Embeddings (required fields)
Workspace — Select the workspace where vectors will be stored.
Collection — Name the collection that will store your embeddings.
Directory — Select a logical folder or namespace inside the collection, or create a new one.
Embedding Model — Choose the embedding model to compute vector representations.
Recommendation: choose a model tuned for your domain, or a general semantic model. Higher-dimension models capture more nuance but cost more.
Full Text Search Index
Search Engine — Select the search engine for lexical (keyword) search.
Lexical Search Storage — Select the index or storage to use for keyword lookups.
Why configure both: semantic search (vector) and lexical search (keyword) together produce the best results — vectors handle meaning; lexical handles exact matches, filters, and facets.
Step 3 — Schedule the Job
Goal: schedule the knowledge graph creation job.
Task Scheduling
Recurring Job — Toggle on and choose the frequency at which the job should run.
Advanced Settings — Set the Degree of Parallelism and Number of Cores for Spark to parallelize data processing.
Task Summary
Review the details.
Click Create Task to submit.
Monitor the Knowledge Graph Creation
The submitted job appears in the Knowledge Graph list alongside existing graphs.
What to Expect in the Resulting Knowledge Graph
Nodes: extracted entities (people, organizations, products, documents, events, and custom types).
Edges: inferred relationships (authored_by, located_in, uses, depends_on, mentions).
Metadata: source, date, confidence score, and provenance (the document or chunk that created the entity).
Capabilities unlocked: semantic search, multi-hop reasoning, graph queries, and dashboards.
Last updated