Step 1
Data
We gather open data (satellite images, weather, statistics, maps) and, where agreed, partner data. For each dataset we record where it came from, when it was updated and what its limits are.
How it works. Four steps, all checkable.
A digital twin is only as good as its data and its honesty about what it does not know. This is how we plan to build them.
Photo: Taylor Vick · Unsplash License01 · The process
From data to a decision.
Our digital twin technology follows four steps: data, twin, scenarios and decisions. At every step we keep notes on sources, assumptions and limits, so the work can be checked by someone else. That way a result is never a black box.
Step 2
The twin
We organise the data into a model of the system: its parts (roads, buildings, people, vegetation, markets) and how they relate. When new observations arrive, the twin updates.
Step 3
Scenarios
We change inputs such as temperature, rainfall, population or demand, and run many simulations. The result is a range of outcomes with probabilities, not a single number.
Step 4
Decisions
We present results in plain language and maps, together with the assumptions, sources and limits of the model. As time passes, we compare forecasts with what actually happened.
02 · The result
A range, not a number.
A scenario run does not produce one forecast but a range of outcomes with probabilities. The chart is illustrative: it shows observations up to today, a likely range that widens with time and a risk threshold. The wider the range, the less sure the model is, and we say so plainly.
- Observed
- Likely range
- Threshold
How to read it
- The black line is what has already been observed.
- The shaded area is the likely range. It widens the further ahead we look.
- The dashed line is a risk threshold set by the user.
- Scenarios A, B, C show how the range moves when assumptions change.
03 · Principles
Principles we follow.
Four rules apply to every twin, whatever its subject. They exist so that every result can be checked, and proved wrong if need be.
01
Ranges, not certainties.
Every forecast comes with its uncertainty.
02
Open sources.
We say where every dataset comes from.
03
Checked against reality.
We measure how well our forecasts did and will publish it once we have enough history.
04
People decide.
The twin supports human judgement; it does not replace it.
04 · Methods
The methods behind the twins.
Digital twin modelling combines several methods, from the agent-based model to scenario analysis. Our articles explain them in plain words, with sources.
Methods
Agent-based models, explained, and how to use them without profiling people
What an agent based model is, from Schelling's segregation model to epidemic and evacuation simulations, and how to build one on aggregate data.
2026-11-04 · 6 min read
Methods
What-if analysis vs scenario analysis: a plain guide
What-if analysis vs scenario analysis: what each one is, when to use which, and simple examples for a municipality and a business.
2026-11-04 · 5 min read
Forecasting
How to read a forecast: when "likely" is a number
Why good forecasts give ranges, how the IPCC turns words like "likely" into probabilities, and what research says about forecasting better.
2026-11-03 · 4 min read
Digital twins
Digital twin examples: from aircraft and factories to cities and the planet
Five real digital twin examples, checked against official sources: NASA's concept, a BMW factory, Virtual Singapore, Helsinki 3D and Destination Earth.
2026-11-04 · 6 min read
05 · Questions
Digital twins. Frequently asked questions.
What a digital twin is, how it works, and what it can and cannot do. If your question is missing, write to us.
What is a digital twin?
A digital twin is a data model of a specific real thing, such as a machine, a building, a city, a forest or the whole Earth. Three features set it apart from a map or a spreadsheet: it mirrors one real system, it stays connected to new observations, and it can be run forward to explore "what if?". Our plain guide explains more.
How does a digital twin work?
Our planned process has four steps. We gather open data and record where each dataset came from. We organise it into a model of the system's parts and how they relate, which updates as new observations arrive. We change inputs and run many simulations to get a range of outcomes. Then we present results with their assumptions, sources and limits, and later compare them with what happened.
What types of digital twins exist?
Twins differ mostly by scale. NASA and the US Air Force described a twin of one specific aircraft, updated with its own history. BMW planned its Debrecen plant as a factory twin. Virtual Singapore and Helsinki 3D are city twins, and the EU's Destination Earth is building a twin of the planet. We compare them in our digital twin examples post.
How is a digital twin used?
To test a change in the model before making it in the real world. A factory tests layouts and robot placement; a city estimates rooftop solar potential, flood impact or heating demand; Destination Earth starts with weather extremes and climate adaptation. For Greece, we plan twins for questions such as wildfire evacuation routes, flood paths, urban heat and population change.
Is Kassandra Labs a live product?
Not yet. We are at concept stage. This site is for people who want to join the waitlist or help shape the first pilots.
Can a digital twin predict the future exactly?
No. A twin shows a range of likely outcomes and how sure we are about them. Anyone who promises a single certain answer about a complex system is overselling.
What data do you use?
We plan to start from open public data, such as European satellite and statistics services, and add partner data only with clear agreements. Every source is documented.
Do you use personal data?
Our models work at the level of areas and groups. We avoid personal data wherever we can, and handle any we do need in line with the GDPR.
Who can become a pilot partner?
Researchers, municipalities, public bodies and businesses with a concrete question and, ideally, some data of their own. Apply through the contact page.
Is this investment advice?
No. Kassandra Labs builds research and scenario tools. Nothing we publish is financial or investment advice.
Early access
Got a question about the future?
Join the waitlist or apply for a pilot.
Concept stage. No product is live yet.