Skip to main content
Back to Use Cases

Case Study — National Digital Twin Programme

AI Ontology Extension Generator — National Digital Twin Programme

Production-ready open-source tool published under the National-Digital-Twin GitHub organisation. Apache License 2.0. Automates ontology generation for UK infrastructure digital twins.

Client

NDTP

Dept. for Business and Trade

Workflow

4-step

guided wizard interface

Input Formats

3

CSV, JSON, RDF/Turtle

The Challenge

The National Digital Twin Programme (NDTP) — a UK Government initiative under the Department for Business and Trade — needed to accelerate digital twin creation across UK infrastructure. Ontology development was a critical bottleneck: manual methods were too slow and required specialist expertise not always available within delivery teams. The programme works alongside the Department for Science Innovation and Technology (DSIT), which identified digital twins as a strategic priority in the UK's AI and data strategy.

The challenge was to automate ontology extension while preserving semantic rigour. The tool had to support heterogeneous data formats and integrate with existing linked data assets. Standards alignment with Government Digital Service (GDS) open data principles and the Cabinet Office data quality framework was required to ensure the outputs were fit for reuse across the public sector. The UK AI Safety Institute has identified ontology consistency as a prerequisite for safe AI deployment in critical national infrastructure.

4-Step Wizard Workflow

StepStageWhat happens
1Data ProfilingAutomatic schema inference from CSV, JSON, and RDF/Turtle inputs.
2Named Entity RecognitionNER pipeline extracts domain entities and relationships from raw data.
3LLM-Assisted GenerationLarge language models propose ontology classes, properties, and axioms.
4Validation and RefinementBuilt-in SHACL validation, visualisation, and iterative human review.

Supported Input Formats

FormatTypical use
CSVTabular data from legacy systems
JSONAPI responses and semi-structured datasets
RDF / TurtleExisting semantic web and linked data assets

Outcome

Delivered a production-ready Streamlit web application. Built-in SHACL validation and visualisation tools allow teams to review and refine generated ontologies before publication. The tool is published on GitHub under the National-Digital-Twin organisation. The National Audit Office has highlighted open-source tooling as a key mechanism for reducing duplication of effort in government digital programmes. The Crown Commercial Service G-Cloud framework recognises open-source AI tools as a cost-effective route to procurement compliance for public sector bodies.

  • Code: Apache License 2.0 — free for public and commercial reuse.
  • Documentation: Open Government Licence v3.0.
  • Actively maintained by the NDTP organisation.

"The ontology extension generator shows how open-source AI tooling can accelerate UK infrastructure data standards. By combining named entity recognition with large language models, we reduced what would have been months of manual ontology work to hours."

— Dr Stylianos Kampakis, Managing Director, Tesseract Academy

View the open-source repository

Published on GitHub under the National-Digital-Twin organisation.

View on GitHub