Knowledge Graph Navigator is an independent reference publication about knowledge graphs, entities and structured data, written for developers and SEOs. It publishes explanatory guides, worked examples, query and markup templates, and annotated reading lists. It is not software, a graph database, a visualization tool, or a commercial platform, and it has nothing to sell or install.

What the site covers

The site is organized as a reference rather than a feed, with one page per topic:

• Core Concepts defines knowledge graphs, entities, triples, and ontologies, and separates them from graph databases and vector databases.
• Languages and Formats covers the syntaxes that carry entity data: JSON-LD, RDFa, Microdata, Open Graph, RDF, SPARQL, OWL and SHACL.
• Ecosystems and Knowledge Bases explains the public graphs that search engines and language models draw on, including the Google Knowledge Graph, Wikidata, DBpedia and schema.org.
• Knowledge Graphs for SEO covers entity SEO, semantic SEO, structured data, and how entities figure in AI search.
• Building Knowledge Graphs walks through modeling, extraction, storage, querying, and GraphRAG.
• Recommended Reading lists verified books, papers, specifications, and courses, and the glossary defines every term the site uses.

Who publishes it

The site is published by Ray Martinez. He is VP of SEO at Archer Education, where he works on organic search and AI search visibility for colleges and universities. He writes for Search Engine Land on schema and entity gaps in AI search, teaches Trends in Media Technologies at the University of Florida, and created Search y Sazón, a community meetup for SEOs, content creators and marketers.

Knowledge Graph Navigator is a personal project. It is not published by or on behalf of Archer Education. More about Ray is at raymartinezseo.com, LinkedIn, and X.

Editorial standards

Every page follows the same rules, and they are the reason the site exists:

• Code is run before it is published. Python examples are executed. SPARQL queries are run against the live endpoint they target, and the page reports what they actually return, including false matches.
• Citations are checked against the source. Books, papers, specifications, and courses are confirmed for title, author, year, and status before they are listed. Anything that cannot be confirmed is left out rather than approximated.
• Documented behavior is kept separate from inference. When a page describes how Google or a language model behaves, it says whether that comes from published documentation, from something externally observable, or from practitioner inference.
• No invented numbers. Pages do not state statistics, rankings, benchmark results or vendor claims that have not been confirmed.
• Tools are compared by capability, not ranked. Comparison pages describe what each product does, confirmed from the vendor’s own documentation, and do not assign scores or an order of preference.

Corrections

Errors are corrected on the page they appear on, and the page’s last-updated date changes when they are. To report a mistake, an outdated fact, or a broken example, use the contact page and name the page and the passage. Reports with a source are acted on fastest.

Contact and social

Send questions, corrections, and suggestions through the contact page. The site is also on Instagram, X and Facebook.