SUDOKN

The Supply and Demand Open Knowledge Network (SUDOKN) is an NSF-funded open knowledge network (NSF Award 2333801) that represents the capabilities of small and medium-sized U.S. manufacturers as an interconnected RDF knowledge graph. It is developed at Arizona State University to improve supplier discovery, capability and capacity analysis, and supply-chain resilience.

Upstream sources

• Manufacturer company websites (primary). The knowledge graph is constructed by web-scraping raw text from individual manufacturer websites and extracting manufacturing triples with an LLM-based ETL pipeline (see the SUDOKN/data-ETL-pipeline repository), following the SUDOKN application ontology. These sources are many individual company sites rather than a single named dataset, so they are represented via SUDOKN’s own primary-source provenance rather than a discrete registry resource.
• NAICS classification (reference). Manufacturer records are classified with the North American Industry Classification System (visible in the graph as NAICSClassifier entries). NAICS is maintained by the U.S. Census Bureau, so this reference influence is recorded via the us-census resource.
• Ontology basis. Semantics come from the Industrial Ontology Foundry (IOF) core and supply-chain ontologies and the Basic Formal Ontology (BFO); these are schema vocabularies rather than data sources.

The project page also displays a “Manufacturing Facilities in the U.S.” map credited to HIFLD, but HIFLD is not documented as a data input to the knowledge graph itself and is therefore not recorded as a source.

Automated Evaluation

• View the automated evaluation: sudokn automated evaluation