Climate Data Narrative
A scroll-driven data visualization that reframes 50 years of climate data as a felt experience rather than a chart — built for a general audience.
The Problem with Climate Charts
Most people have seen the hockey-stick graph. Almost no one has changed their behaviour because of it. Data visualization for climate typically produces accurate representations of the problem and zero emotional resonance. The numbers are right; the medium is wrong.
This project asked: what would it take to make climate data felt?
The Approach
I built a scroll-driven narrative that takes 50 years of NOAA temperature anomaly data and translates it into a sensory experience. No axes, no labels, no colour-coded legends.
The visual language: a slow, evolving shader landscape. As the user scrolls forward in time, the landscape becomes hotter — not through numbers, but through the quality of light, the frequency of tremor in the terrain, the desaturation of colour at the periphery. The progression is perceptible before it is legible.
Technical architecture:
- React scroll context tracking viewport position → normalized time value (0–1)
- Three.js terrain mesh with 50k vertices, displaced by NOAA data mapped to height
- Custom GLSL shader reading temperature anomaly as a uniform → drives colour temperature
- Framer Motion for the sparse text overlays that provide the only explicit data context
Outcome
Published as a standalone URL. Visited by 40,000+ people in the first month, primarily via climate and design communities. Zero budget. The project was used as a teaching reference in two university data visualization courses.
The lesson I took from this: the constraint of removing traditional chart language forces you to think about what the data means, not just what it says.