Knowledge Graph Building from NLP using Spacy

---

A simple example of producing a knowledge graph using NLP the example uses the following tooling:

• Jupyter Labs and Python
• Spacy NLP
• Networkx

In this example I take a low effort approach by using off-the-shelf models and tooling. I wanted to evaluate the default models in Spacy to process a document with the following processing pipeline:

1. Named Entity Extraction NER
2. Co-reference resolution
3. Semantic type enhancement using wikidata
4. Relationship extraction
5. Create NetworkX graph

From a limited set of control data and a small amount of code the output is reasonable. It’s a simple model so could be extended View Source in Html

Document Input

Knowledge Graph Output