SPARQL Query Examples

Use these SPARQL queries to analyze ORP documents as linked data. All queries work on ORP JSON-LD documents loaded into a triple store.

What is SPARQL?

SPARQL is the query language for RDF data (like SQL for semantic web). When you convert ORP documents to JSON-LD and load them into a triple store, you can ask complex questions across all your documents.

Try it yourself:

Load ORP JSON-LD documents into Apache Jena Fuseki
Copy queries from this page
Paste into Fuseki’s query interface
See results instantly

Basic Queries

Query 1: List All Documents

PREFIX orp: <https://openreason.org/vocab/>
PREFIX dc: <http://purl.org/dc/terms/>
 
SELECT ?doc ?title ?created
WHERE {
  ?doc a orp:Document ;
       dc:title ?title ;
       dc:created ?created .
}
ORDER BY DESC(?created)

What it finds: All ORP documents with their titles and creation dates, sorted by newest first.

Use case: Document discovery and inventory.

Query 2: Find Document Authors

PREFIX dc: <http://purl.org/dc/terms/>
PREFIX schema: <http://schema.org/>
 
SELECT ?doc ?title ?authorName ?role
WHERE {
  ?doc dc:title ?title ;
       dc:creator ?author .
  ?author schema:name ?authorName .
  OPTIONAL { ?author orp:role ?role }
}

What it finds: All documents with their authors and roles.

Use case: Attribution and responsibility tracking.

Layer 1: Data Provenance Queries

Query 3: Find All Datasets

PREFIX orp: <https://openreason.org/vocab/>
PREFIX dcat: <http://www.w3.org/ns/dcat#>
PREFIX dc: <http://purl.org/dc/terms/>
 
SELECT ?dataset ?name ?description ?steward
WHERE {
  ?doc orp:hasLayer1 ?l1 .
  ?l1 orp:dataset ?dataset .
  ?dataset dcat:title ?name ;
           dc:description ?description ;
           prov:wasAttributedTo ?steward .
}

What it finds: All datasets mentioned across all ORP documents.

Use case: Data catalog creation.

Query 4: Datasets by Funding Source

PREFIX orp: <https://openreason.org/vocab/>
PREFIX schema: <http://schema.org/>
 
SELECT ?dataset ?name ?funder ?amount
WHERE {
  ?doc orp:hasLayer1 ?l1 .
  ?l1 orp:dataset ?dataset .
  ?dataset dcat:title ?name ;
           orp:fundingSource ?funding .
  ?funding schema:name ?funder .
  OPTIONAL { ?funding orp:amount ?amount }
}
ORDER BY DESC(?amount)

What it finds: Datasets sorted by funding amount, showing who funded them.

Use case: Track funding influence on research.

Query 5: Datasets with Exclusions

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT ?doc ?dataset ?name ?exclusion ?rationale
WHERE {
  ?doc orp:hasLayer1 ?l1 .
  ?l1 orp:dataset ?dataset .
  ?dataset dcat:title ?name ;
           orp:exclusionCriterion ?exclusion .
  OPTIONAL { ?exclusion orp:rationale ?rationale }
}

What it finds: All datasets that excluded data, with documented rationale.

Use case: Identify potential selection bias.

Query 6: Temporal Coverage Analysis

PREFIX dcat: <http://www.w3.org/ns/dcat#>
 
SELECT ?dataset ?name ?start ?end
WHERE {
  ?dataset dcat:title ?name ;
           dcat:temporalCoverage ?temporal .
  ?temporal dcat:startDate ?start ;
            dcat:endDate ?end .
}
ORDER BY ?start

What it finds: Date ranges covered by each dataset.

Use case: Identify data recency and coverage gaps.

Layer 3: Empathy Mapping Queries

Query 7: Stakeholders with No Representation

PREFIX orp: <https://openreason.org/vocab/>
PREFIX schema: <http://schema.org/>
 
SELECT ?doc ?stakeholder ?name ?population
WHERE {
  ?doc orp:hasLayer3 ?l3 .
  ?l3 orp:stakeholder ?stakeholder .
  ?stakeholder schema:name ?name ;
               orp:representation "None" ;
               orp:estimatedPopulation ?population .
}
ORDER BY DESC(?population)

What it finds: “Absent nodes” - stakeholder groups with zero voice in decisions, sorted by population affected.

Use case: Critical governance gap analysis.

Query 8: Stakeholders by Net Welfare Impact

PREFIX orp: <https://openreason.org/vocab/>
PREFIX schema: <http://schema.org/>
 
SELECT ?doc ?stakeholder ?name ?impact ?population
WHERE {
  ?doc orp:hasLayer3 ?l3 .
  ?l3 orp:stakeholder ?stakeholder .
  ?stakeholder schema:name ?name ;
               orp:netWelfareImpact ?impact ;
               orp:estimatedPopulation ?population .
  FILTER(?impact < 0)
}
ORDER BY ?impact

What it finds: Stakeholders experiencing net negative impact, sorted by severity.

Use case: Identify populations bearing disproportionate harm.

Query 9: Minority Stakeholder Representation

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT ?doc ?stakeholder ?name ?representation ?population
WHERE {
  ?stakeholder orp:representation ?representation ;
               orp:estimatedPopulation ?population ;
               schema:name ?name .
  FILTER(?population < 1000000)
}
ORDER BY ?population

What it finds: Small stakeholder groups and their governance representation.

Use case: Stress-test minority protection (REE Principle 4: Universal Sentience).

Layer 4: Accountability Queries

Query 10: Decision Audit Trail

PREFIX orp: <https://openreason.org/vocab/>
PREFIX prov: <http://www.w3.org/ns/prov#>
 
SELECT ?doc ?decision ?timestamp ?decisionMaker ?logic
WHERE {
  ?doc orp:hasLayer4 ?l4 .
  ?l4 orp:decision ?decision .
  ?decision prov:atTime ?timestamp ;
            prov:wasAssociatedWith ?decisionMaker ;
            orp:constitutiveLogic ?logic .
}
ORDER BY ?timestamp

What it finds: Chronological log of who decided what and why.

Use case: Accountability audit (REE Principle 5: Transparent Accountability).

Query 11: Decisions by Type

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT ?type (COUNT(?decision) as ?count)
WHERE {
  ?doc orp:hasLayer4 ?l4 .
  ?l4 orp:decision ?decision .
  ?decision orp:decisionType ?type .
}
GROUP BY ?type
ORDER BY DESC(?count)

What it finds: Distribution of decision types across documents.

Use case: Identify common decision patterns.

Layer 5: Fork Analysis Queries

Query 12: Fork Genealogy

PREFIX orp: <https://openreason.org/vocab/>
PREFIX dc: <http://purl.org/dc/terms/>
 
SELECT ?fork ?forkTitle ?original ?type ?description
WHERE {
  ?fork orp:forkedFrom ?original ;
        orp:forkType ?type ;
        dc:title ?forkTitle ;
        dc:description ?description .
}

What it finds: All alternative proposals and challenges, showing what they forked from.

Use case: Track reasoning evolution through critique and alternatives.

Query 13: Most Challenged Documents

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT ?original (COUNT(?fork) as ?challengeCount)
WHERE {
  ?fork orp:forkedFrom ?original ;
        orp:forkType "challenge" .
}
GROUP BY ?original
ORDER BY DESC(?challengeCount)

What it finds: Documents with the most challenges, indicating contentious reasoning.

Use case: Identify high-controversy decisions requiring public debate.

Cross-Document Pattern Queries

Query 14: Pharmaceutical Funding Across All Documents

PREFIX orp: <https://openreason.org/vocab/>
PREFIX schema: <http://schema.org/>
 
SELECT ?doc ?dataset ?funder ?amount
WHERE {
  ?doc orp:hasLayer1 ?l1 .
  ?l1 orp:dataset ?dataset .
  ?dataset orp:fundingSource ?funding .
  ?funding schema:name ?funder .
  OPTIONAL { ?funding orp:amount ?amount }
  FILTER(CONTAINS(LCASE(?funder), "pharma"))
}

What it finds: All datasets funded by pharmaceutical companies across your entire document collection.

Use case: Systematic conflict-of-interest analysis.

Query 15: Documents Affecting Specific Population

PREFIX orp: <https://openreason.org/vocab/>
PREFIX schema: <http://schema.org/>
 
SELECT DISTINCT ?doc ?title
WHERE {
  ?doc dc:title ?title ;
       orp:hasLayer3 ?l3 .
  ?l3 orp:stakeholder ?stakeholder .
  ?stakeholder schema:name ?name .
  FILTER(CONTAINS(LCASE(?name), "renters"))
}

What it finds: All policy documents that affect a specific stakeholder group (e.g., renters).

Use case: Stakeholder-centric policy review.

Query 16: AI Training Datasets (Post-Hoc Reconstructions)

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT ?doc ?title ?extension
WHERE {
  ?doc dc:title ?title ;
       orp:postHoc true ;
       orp:extension ?extension .
  FILTER(CONTAINS(?extension, "ai-training"))
}

What it finds: All post-hoc reconstructions of AI training datasets.

Use case: Discover accountability gap documentation in AI systems.

Advanced Analytics

Query 17: Absent Node Population Impact

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT (SUM(?population) as ?totalAffected)
WHERE {
  ?stakeholder orp:representation "None" ;
               orp:estimatedPopulation ?population .
}

What it finds: Total number of people affected by decisions with zero representation.

Use case: Quantify governance gap at portfolio level.

Query 18: Compliance Level Distribution

PREFIX orp: <https://openreason.org/vocab/>
 
SELECT ?level (COUNT(?doc) as ?count)
WHERE {
  ?doc orp:complianceLevel ?level .
}
GROUP BY ?level
ORDER BY ?level

What it finds: How many documents achieve each compliance level (ORP-Basic, ORP-Standard, ORP-Full, ORP-PostHoc).

Use case: Track documentation quality across portfolio.

Performance Tips

Optimize Queries

Use LIMIT for large datasets:

SELECT ?doc ?title WHERE { ... } LIMIT 100

Filter early in the query:

WHERE {
  ?doc orp:complianceLevel "full" .  # Filter first
  ?doc dc:title ?title .
}

Index frequently queried fields in your triple store
Use OPTIONAL sparingly - it’s expensive

Try These Queries

Setup Fuseki Triple Store

# Download and start Fuseki
wget https://dlcdn.apache.org/jena/binaries/apache-jena-fuseki-4.9.0.tar.gz
tar xzf apache-jena-fuseki-4.9.0.tar.gz
cd apache-jena-fuseki-4.9.0
./fuseki-server --update --mem /orp
 
# Load ORP JSON-LD document
curl -X POST \
  -H "Content-Type: application/ld+json" \
  --data-binary @danish_property_tax_reform.jsonld \
  http://localhost:3030/orp/data
 
# Visit http://localhost:3030/
# Click "Query" tab
# Paste queries from this page

Python SPARQL Example

from rdflib import Graph
 
# Load ORP document
g = Graph()
g.parse("danish_property_tax_reform.jsonld", format="json-ld")
 
# Run Query 7: Absent nodes
query = """
PREFIX orp: <https://openreason.org/vocab/>
PREFIX schema: <http://schema.org/>
 
SELECT ?name ?population WHERE {
  ?stakeholder schema:name ?name ;
               orp:representation "None" ;
               orp:estimatedPopulation ?population .
}
ORDER BY DESC(?population)
"""
 
for row in g.query(query):
    print(f"{row.name}: {row.population:,} people")

Next Steps

Try JSON-LD - See JSON-LD Format guide
Install Fuseki - Set up your own SPARQL endpoint
Load Documents - Convert your ORP YAML to JSON-LD
Run Queries - Copy examples from this page

More Queries: See sparql_examples.md in the repository for 15+ additional examples with detailed explanations.

JSON-LD Format Overview