Communicating urban heat to a non-specialist audience means communicating a simulation: radiation values over a built environment, varying by time of day, changing under different mitigation strategies. That is three dimensions of variation and a lot of numbers, and it is normally delivered as a slide.
Climate Physicalization, from Milena Vuckovic, Marlene Huber, Fabrizia Bando-Bechtold, Daniel Cornel and Daniel Pahr, is a situated geospatial thermal physicalization — a physical object you handle instead.
What the thing is
Three components, and the design logic runs through all of them:
A 3D model of an urban neighbourhood, with radiation data represented as a layered contour structure. Contours are the right encoding for a physical build: a stack of layers is cheap to fabricate, reads as a surface at a glance, and gives a finger something to trace. You can feel where the gradient is steep.
Three interchangeable climate adaptation scenarios. This is the part that makes it an interface rather than a sculpture. A scenario — more tree cover, a changed surface material, an altered building arrangement — is a physical part you lift out and replace. Comparing two futures is a swap you perform with your hands, not a toggle in a UI.
An externally mounted, electronically driven solar path, with multiple discrete LEDs that can be switched on and off at will. The sun’s trajectory across the sky, rendered as a rail of individually addressable lights.
That last element is the cleverest piece. Solar radiation on a city block is a function of sun position, so the independent variable in the whole simulation is “where is the sun”. Putting that on a physical arc with discrete LEDs makes the variable manipulable and legible simultaneously — you see which position is selected because it is lit, and you change it by selecting another. No legend, no clock, no mental conversion from a time-of-day slider to a solar angle.
What they found
A hands-on qualitative lab study with 16 participants, using think-aloud, observational and questionnaire protocols.
Participants “responded very favorably to the model,” highlighting the novelty of the visualization approach and its capacity to improve the interpretability of the presented information.
Interpretability is the claim worth taking seriously, and it is a narrower claim than “engagement”. A physicalization that people enjoy handling is easy to achieve. One where handling it makes the data clearer is the actual research question, because the failure mode of this genre is a beautiful object that communicates less than the chart it replaced.
Being precise about the evidence: this is a qualitative study with 16 participants using think-aloud and questionnaires. It reports that people found the model interpretable and liked it. It does not measure comprehension against a baseline visualisation, so “clearer than a map” is not what has been shown. The paper is ten pages and frames itself as introducing a prototype, which is the right scope.
Why it belongs in this topic
The lineage here is direct. Data physicalization descends from the tangible-interaction research programme — the argument that physical form carries meaning that screen representations cannot, and that the hands are an underused input channel. That is the position Hiroshi Ishii has held at the MIT Media Lab for three decades, and which won him an award last week while the consumer world went the other way.
This paper is that argument applied to a case where it is unusually defensible. Urban heat is inherently spatial and inherently three-dimensional — shadows, building geometry, surface materials, sun angle. Flattening it to a screen throws away a dimension the data actually has. A physical model does not have to simulate depth because it has depth.
It also sidesteps the standard objection to physicalization, which is that a physical artefact cannot be re-rendered for new data. Here the data is fixed — it is a neighbourhood — and what varies is the scenario, which is exactly what has been made swappable. The design works because the authors put the mutability where the question is.
What a practitioner can take from it
For anyone building tangible or exhibition work, three transferable moves:
Encode the field as layers, not as height. A contour stack is fabricable, readable and touchable, and it tolerates imprecision in the fabrication better than a smooth surface does.
Make the comparison physical. Swapping a part is a stronger comparison action than pressing a button, because the old state is in your other hand. Visitors understand “these are alternatives” without being told.
Put the independent variable on a rail of lights. Discrete, addressable, individually switchable positions are legible, self-documenting and robust. There is no state hidden from the visitor, and nothing to reset when the gallery opens.
That last point is the quiet virtue of this whole approach. A physicalization with interchangeable parts and an LED array has very little that can go wrong overnight — which is more than can be said for most installations built on a sensor and a projector.
Sources
- Climate Physicalization: Bridging Tangible Interaction and Data Visualization for Environmental Communication — arXiv:2610.21401
- DOI: 10.48550/arXiv.2610.21401
- Data physicalization — research community overview
- Tangible Media group — MIT Media Lab