Connect
Different evidence.
A shared representation.
Bring Earth observation, GIS, administrative records, indicators, policy, text, and local observations into a common analytical foundation.
Capabilities / Purpose into practice
Understand the city.
Build on the evidence.
Our tools begin with a shared purpose: helping people make better decisions about the built environment. City foundation models connect physical conditions with the policies, histories, and local knowledge that give a place meaning.
A connected system
01 / City foundation models
Foundation models,
not digital twins.
The goal is a useful representation of a city. Not a replica of everything in it.
Our models are designed to learn relationships among buildings, climate, policy, and local conditions. They connect physical patterns with things that are related by meaning, purpose, or everyday experience.
Each model is locally designed. Its evidence, methods, and questions reflect a city’s particular physical environment, social fabric, and administrative needs.
Different evidence.
Understand
Relationships beyond
the map.
Explore connections between physical characteristics and the less visible conditions that shape life in a place.
Apply
Local detail.
Practical questions.
Work at the scale of parcels, buildings, neighbourhoods, and municipal decisions, while retaining the context around them.
02 / Forge
The system behind
the intelligence.
Forge builds the models—and develops the methods that make the next model better.
Our internal knowledge and development system connects agents, evidence, analytical knowledge, and accumulated experience. It gives model development a foundation that grows with the work.
- Organise the knowledge. Bring research, methods, and experience into the development process.
- Develop and test. Use agents and analyst expertise to investigate evidence and evaluate approaches.
- Improve the methodology. Carry learning forward into the construction and updating of city models.
As Forge improves, updates can become increasingly autonomous. Analysts remain in the loop, and changes remain connected to their evidence and methodology.
03 / Core
Put models to work.
Keep the evidence.
Core serves city foundation models alongside BIT’s global and country-level indicators.
It is the operational interface for accessing and maintaining these resources. Core is deliberately evidence-preserving: it does not independently improve itself or replace the evidence beneath a result.
- Lineage. Connect outputs to the data and transformations behind them.
- Provenance. Retain where evidence came from and how it was used.
- Reproducibility. Maintain methods, versions, and the underlying evidence.
Updates are traceable, with analysts responsible for oversight and interpretation.
Whole-building analysis
A building,
inside and out.
In development / Pilot testing
Connect the view from above with the spaces within.
We are developing an integrated approach to interior and exterior building analysis. Earth observation (EO) provides exterior and neighbourhood context; our ML-powered, privacy-protective LiDAR app brings premises scanning into the same evidence base.
The aim is a more complete understanding of a building: its physical characteristics, interior spaces, and relationship to the place around it. That evidence can support retrofit planning, property analysis, and richer city foundation models.
Respect for occupants’ privacy guides the scanning app’s development. Machine learning helps turn spatial observations into evidence that analysts can review and interpret.
Earth observation
Building exteriors and the surrounding environment.
LiDAR premises scanning
Interior spatial evidence, with privacy guiding the design.
03 / Connected understanding
One building.
A fuller picture.
Machine learning and analyst interpretation bring exterior context and interior detail together.
Conceptual approach / Pilot capabilityFrom individual premises to neighbourhoods, deeper observation strengthens the evidence for better buildings.
Discuss a pilot
Experimental / Kitchener–Waterloo
Mariposa.
Where we learn
what a city can tell us.
Our experimental city foundation model is grounded in Kitchener–Waterloo.
Its name is inspired by Stephen Leacock’s Mariposa: a small place in which, in a sense, everything happens. It is our testing ground for locally designed intelligence.
Mariposa is an active experiment.
Discuss a partnershipOur strategic direction
Local understanding.
A wider perspective.
A growing portfolio of deeply observed cities can strengthen the way we measure the built environment at every scale.
Global foundation models provide valuable general representations. Our city-scale approach adds the local detail needed for analysis and administration.
Our research foundationsBuilding Insights Together