Anthroposense

Making climate recommendations visible inside the room they affect.

Anthroposense turns climate advice into spatial tasks using LLMs, a digital twin of the room, and VR. In the headset, a recommendation becomes a concrete prompt: move the side table, shift the sofa, let in more daylight. The carbon and engagement scores react as you go.

It ends with a final carbon score for the room. A plan view marks what moved, so an abstract sustainability number reads as a concrete before-and-after.

It starts with a scan. Users capture their room with a LiDAR app, which builds the digital twin the VR recommendations run on.

Inside the twin, the visualization layer explains why something should move. Tooltips tie a specific table position to daylight, ventilation, artificial lighting, and air-conditioning load.

As the user follows the tooltips, the carbon score updates live. That loop is why the project is interactive at all: climate impact lands harder when every move changes the number in front of you.

The prototype ran on Oculus, Microsoft MRTK, Unity, and a few social APIs. Testing in the user's real room, not a generic scene, kept the experience grounded.