README.md

metadata

language:
  - en
license: apache-2.0
size_categories:
  - 1K<n<10K
task_categories:
  - text-retrieval
  - question-answering
  - zero-shot-classification
tags:
  - legal
  - gdpr
  - privacy
  - compliance
  - knowledge-graph
  - normative-triples
  - shacl
  - provenance
  - knowledge-engineering
  - symbolic-ai
  - privacy-engineering
pretty_name: GDPR Normative Triples & Compliance Knowledge Graph

GDPR Normative Triples & Knowledge Graph Benchmark Dataset

A comprehensive, deterministic, cryptographically provenanced legal knowledge engineering dataset encoding the full normative and relational structure of Regulation (EU) 2016/679 (General Data Protection Regulation - GDPR).

---

🔬 Dataset Overview & Knowledge Engineering Rigor

In legal informatics and regulatory AI, relying on ungrounded language models introduces significant risks of hallucinated citations, missing conditions, or conflated legal roles. This dataset provides the first complete, line-by-line, peer-verifiable ontological decomposition of the GDPR into:

• 936 Validated Normative Triples with Deontic Logic Modal Classification.
• 3,212 Semantic Knowledge Graph Entities and 4,500+ Relational Edges.
• 1,000+ Controlled Vocabulary Terms across 21 domain classes.
• W3C SHACL Constraints & First-Order Compliance Rules for executable symbolic validation.
• Cryptographic 4-Tier SHA-256 Provenance Ledger linking every triple to EUR-Lex CELEX:32016R0679.

---

📂 Repository & Asset Structure

1. gdpr_normative_triples.json

Full Subject-Predicate-Object decomposition of all 99 Articles across all 11 Chapters:

{
  "id": "TRIPLE-ART-0033-P1-1",
  "article_number": 33,
  "paragraph_number": "1",
  "subject": "Data Controller",
  "modality": "OBLIGATION",
  "predicate": "shall notify personal data breach to competent supervisory authority without undue delay and where feasible not later than 72 hours",
  "object": "Supervisory Authority",
  "source_text": "In the case of a personal data breach, the controller shall without undue delay and, where feasible, not later than 72 hours after having become aware of it, notify the personal data breach to the supervisory authority competent in accordance with Article 55...",
  "sha256_hash": "a4b89d6e492e86bf4148bcf7eb76ec965b6f3b0e36398b18ec4908ef484be49d"
}

2. gdpr_knowledge_graph.json

Graph network topology ready for Cytoscape.js, NetworkX, Neo4j, or RDF triple stores:

• elements.nodes (3,212 entities): Categorized by type (actor, obligation, right, article, authority, penalty).
• elements.edges (4,500+ edges): Classified by predicate relation (governs, requires, exempts, reports_to, imposes_fine).

3. domain_data_dictionary.json

21 distinct ontological domain classes including:

• Legal Actors & Statutory Roles
• Processing Operations & Lawful Bases (Article 6)
• Data Subject Rights (Articles 12–22)
• Technical and Organizational Security Measures (TOMs - Article 32)
• Statutory Fines & Penalty Tiers (Article 83)
• Cross-Border Transfer Mechanisms (Chapter V)

4. provenance_ledger.json

Complete cryptographic audit ledger containing SHA-256 hashes of the canonical EUR-Lex text, individual article paragraphs, and derived knowledge triples.

5. rules/ (Executable Knowledge Engineering Modules)

• gdpr_shacl_shapes.json & gdpr_shacl_shapes.ttl: W3C SHACL constraint shapes for Article 30 ROPAs and Article 37 DPOs.
• edpb_fine_guidelines.json: Formalized EDPB Guidelines 04/2022 penalty calculation matrix.
• privacy_lint_rules.json: DevSecOps static analysis rules for Terraform and OpenAPI.
• dpia_guidelines_wp248.json: EDPB WP 248 rev.01 Article 35 high-risk screening criteria.
• dsar_rules.json: Article 12(3) SLA deadlines and Article 17(3) statutory exemption rules.
• cross_regulatory_frameworks.json: Bidirectional cross-walks to ISO/IEC 27701:2019, CCPA/CPRA, and HIPAA.

---

🐍 Usage in Python

Querying Normative Triples

import json

# Load normative triples
with open('gdpr_normative_triples.json', 'r', encoding='utf-8') as f:
    triples = json.load(f)

# Filter all strict prohibitions for Data Processors
prohibitions = [
    t for t in triples 
    if t['modality'] == 'PROHIBITION' and 'Processor' in t['subject']
]
print(f"Total Processor Prohibitions: {len(prohibitions)}")

Loading Knowledge Graph in NetworkX

import json
import networkx as nx

with open('gdpr_knowledge_graph.json', 'r', encoding='utf-8') as f:
    kg = json.load(f)

G = nx.DiGraph()
for node in kg['elements']['nodes']:
    G.add_node(node['data']['id'], label=node['data']['label'], type=node['data'].get('type'))

for edge in kg['elements']['edges']:
    G.add_edge(edge['data']['source'], edge['data']['target'], relation=edge['data']['label'])

print(f"Knowledge Graph: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges")

---

🎯 Downstream Applications & Research Tasks

• Neuro-Symbolic Legal Reasoning: Grounding LLM responses in deterministic deontic triples to eliminate hallucinations.
• Automated Regulatory Compliance Checking: Validating ROPA databases and cloud architectures with W3C SHACL shapes.
• Cross-Jurisdictional Harmonization: Translating compliance proofs between GDPR, ISO/IEC 27701, CCPA, and HIPAA.
• Privacy-as-Code DevSecOps Integration: Continuous integration scanning of cloud infrastructure for privacy debt.

---

📜 Legal Provenance & Citation

@dataset{gdpr_normative_triples_2026,
  author    = {Mujtaba Hussain},
  title     = {GDPR Normative Triples and Compliance Knowledge Graph},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/gitmodelmujtaba/gdpr-normative-triples}
}

Derived from Regulation (EU) 2016/679 of the European Parliament and of the Council (OJ L 119, 4.5.2016). Licensed under Apache 2.0.