Wilt: five solar-powered paper flowers that bloom or struggle on live air-quality data

For a final-year project at Queen Mary University of London, Dharma Rumpal set out to test whether interactive art powered by renewable energy could engage people on climate change in a way that data, scientific reports and headlines never could.

Five servo-driven paper flowers breathe with live air-quality data from ten rotating world cities — clean air opens the blooms, polluted air slows them to a struggle.

An Arduino Uno drives five SG90 servos through a synchronised breathing function, with speed and range calibrated to each air-quality state; a Python script polls the OpenWeatherMap Air Pollution API and cycles through ten cities every 45 seconds, from Reykjavik to Delhi and Beijing. A visible 5W solar panel charges an 18650 battery — the rig deliberately left exposed because the power system is part of the message. A companion site synced a live city monitor and a global air-quality map with the flowers, and Rumpal's brother Cosmo composed the music.

Evaluated at a QMUL exhibition with paired pre/post surveys and interviews: climate concern rose from 3.84 to 4.48 out of 5, motivation to act from 3.56 to 4.35, and the share who considered interactive art an effective way to communicate climate change jumped from 20% to 96%. The dominant emotional responses were curiosity, hope and inspiration, not fear or helplessness.

Wilt is a solar-powered kinetic art installation built by Dharma Rumpal as a final-year project at Queen Mary University of London. Five servo-driven paper flowers breathe in response to live air-quality data fetched from cities around the world, cycling through ten of them every 45 seconds — from Reykjavik, described as one of the cleanest cities on earth, to Delhi and Beijing at the other extreme. Good air quality produces full, open blooms moving freely; poor air quality produces slow, laboured, restricted movement.

The build layers consumer hardware and open APIs: an Arduino Uno manages the five servos, a Python script makes the live calls to the OpenWeatherMap Air Pollution API and sends commands over USB serial, and a 5W solar panel charges an 18650 lithium-ion cell. The solar rig was deliberately left visible, because the power system is part of the message rather than infrastructure to hide. A companion site extended the piece with a live city monitor kept in sync with the flowers and an interactive global air-quality map.

What Rumpal argues separates Wilt from much earlier climate art is that it was systematically evaluated. Exhibition participants completed paired pre- and post-surveys and semi-structured interviews: climate concern rose from 3.84 to 4.48 out of 5, motivation to act from 3.56 to 4.35, and belief in interactive art as an effective climate-communication method went from 20% before the exhibition to 96% after it. The dominant emotional responses were curiosity, hope and inspiration rather than fear or helplessness.

The power source is part of the artwork: the solar rig is deliberately left visible instead of hidden.

It maps an abstract global dataset onto a body — flowers that visibly struggle when the air is bad.

It was systematically evaluated with paired pre/post surveys, which the writeup says most climate art skips.

The measured emotional response was curiosity, hope and inspiration rather than fear.

Make the data's cause physical — the visible solar rig carries the message as much as the flowers do. And evaluate: paired surveys turn 'people seemed engaged' into reusable evidence.

Rumpal writes that a methodological template now exists for combining data-responsive kinetic art, renewable-energy infrastructure and rigorous evaluation, one that future artists, researchers and climate communicators can take, replicate and build upon. The writeup appeared on the Sensing the Forest research blog on 1 October 2026.