UN-GGCE Deliverables Portal

Appearance

Return to top
On this page

UN-GGCE Context Graph Architecture - Kickoff with Kurt Cagle

Project Summary

The United Nations Global Geodetic Centre of Excellence (UN-GGCE) is tasked with mapping and assessing the Global Geodesy Supply Chain. Currently, the maturity assessment and supply chain data are managed across spreadsheets (e.g., Work Assignment 3...xlsx) and a relational database (geodetic_maturity.db).

To address the UN's rigorous requirements for traceability, provenance, and complex relationship mapping (linking capabilities, observatories, products, and value streams), we are transitioning to a Context Graph Architecture grounded in semantic web standards. We have established an initial ontology (ggsc-ontology.ttl) leveraging BACM (Business Architecture Core Metamodel), GeoSPARQL, and OWL/RDF.

The objective of this project is to implement a robust, scalable knowledge graph that enables dynamic analysis of supply chain risks, maturity, and critical paths, facilitating data-driven investment decisions.

Objectives for the Kickoff

  • Align on the vision for the Context Graph Architecture for the UN-GGCE.
  • Review the current state of the ontologies and data sources.
  • Define the technical stack and next steps for the graph implementation.

Roles & Division of Labor

Lead UN Consultant (Ben) — The "What" and the "Why"

Focus: Project Requirements, Domain Knowledge, and Final Deliverables.

  • Requirements & Mandates: Own the project requirements based on the UN Terms of Reference. Define what questions the graph must answer (e.g., EGV traceability, maturity gaps).
  • Domain Knowledge Translation: Act as the bridge between geodetic domain experts and the technical architecture. Provide raw data (geodetic_maturity.db, spreadsheets) and validate graph accuracy.
  • Stakeholder Management & Reporting: Lead the integration of graph insights into final consultancy outputs (e.g., "State of Geodesy 2026").
  • Project Governance: Oversee YouTrack and ensure the timeline aligns with UN milestones.

Principal Architect & Lead Ontologist (Kurt Cagle) — The "How"

Focus: Structural Design, Standards Alignment, and Technical Strategy.

  • Ontology Design & W3C Alignment: Lead the structural design of the ontology (RDF-star, SHACL). Ensure alignment with W3C standards and Decision Trace methodologies.
  • Architecture Strategy: Define the approach for managing multiple graphs (Agent personas vs. UN-GGCE project data vs. InfoSec) and retrospective data implementation.
  • Infrastructure & Tooling Guidance: Recommend the optimal graph database/triple store (GCP, OVHcloud, etc.) and development workflows (Git).
  • Technical Mentorship: Guide the ETL strategy for converting existing relational/tabular data into the new graph structure.

Proposed Task List & Discussion Points

Phase 1: Ontology Review & Validation

  • [ ] Review Current .ttl Files: Walk through ggsc-ontology.ttl, ggsc-capabilities.ttl, ggsc-observatories.ttl, etc.
  • [ ] Assess BACM Integration: Validate our use of the Business Architecture Core Metamodel (BACM) for modeling capabilities, organizations, and value streams.
  • [ ] Data Granularity: Discuss the modeling of Essential Geodetic Variables (EGVs) and Level 1/2/3 capabilities.
  • [ ] Provenance & Traceability: Ensure the ontology supports UN traceability requirements (e.g., tracking data origins, processing levels, and organizational accountability).

Phase 2: Data Ingestion & Graph Setup

  • [ ] Graph Database Selection: Evaluate and select a target triple store / graph database (e.g., GraphDB, Stardog, Neptune, or open-source alternatives).
  • [ ] ETL Pipelines: Design the process to migrate existing tabular data (geodetic_maturity.db, Excel) into RDF triples.
  • [ ] Automation: Review existing Python scripts (src/scripts/maturity_sql_import.py) and determine how to refactor them for RDF generation (e.g., using rdflib).

Phase 3: Querying & Visualization

  • [ ] SPARQL Endpoints: Define the core SPARQL queries needed to answer key supply chain risk and maturity questions (e.g., "Which Level 3 capabilities are critically under-matured?").
  • [ ] Geospatial Queries: Discuss leveraging GeoSPARQL for observatory mapping and geographic redundancy analysis.
  • [ ] Visualizing the Graph: Plan the integration of graph visualization tools or generating Mermaid.js diagrams directly from graph queries to support reports and presentations.
  • [ ] Presentation Layer Technology: Discuss the stack for the visual presentation layer.
    • [ ] How to render the Capability Model with dynamic Maturity Layers (People, Process, Technology, Data).
    • [ ] Strategy for presenting EGV Workflows (e.g., dynamic SVG, Mermaid integration, or custom D3.js visualizations from graph data).

Phase 4: Roadmap & Governance

  • [ ] Continuous Integration: Establish governance for ontology updates and data quality assurance.
  • [ ] Integration with Reporting: Plan how the graph will feed into the final consultancy outputs (e.g., "State of Geodesy 2026").
  • [ ] Milestones: Finalize the timeline and responsibilities for the knowledge graph rollout.

Meeting: 19 Mar 26 - Logistical & Technical Deep Dive

Availability & Cadence

  • Commitment: How many days per week would you need/likely be able to work? (Focused mostly on the graph design—RDF-star, SHACL—initially?)
  • Syncing: How often would it make sense to connect?

Architecture & Security

  • Multi-Graph Management: How do we handle multiple graphs (e.g., Agent personas, UN-GGCE project data, InfoSec/Sensitive data)?
  • Retrospective Implementation: Can we implement the architecture retrospectively to existing data/work?
  • Infrastructure: Proposing GCP Cloud Run for hosting remote MCP servers (to handle persistent SSE connections securely), reserving the local machine purely for heavy RDF-star/SHACL processing workloads. Does this align with your preferred architecture?

Tooling & Governance

  • Project Management: Are you happy to work within a project system like YouTrack? (Note: We have an MCP server for integration).
  • Version Control: Do I use a GIT repo for the ontology and graph configuration?
  • Presentation Layer: What technology do we use for the presentation layer? (Requirements: Capability model with maturity layers, EGV workflows).

Meeting Notes & Action Items:(To be filled during the meeting)