GIS-processed satellite data, generative drawing machine on wall
95″ × 106″

Two iterations are shown in this documentation. Both were produced by the same drawing machine, but at different times and using different GIS-processed satellite datasets. The change in data generates new movement patterns, allowing the machine to reinterpret the city with each run. Each iteration becomes its own cartographic event shaped by changing urban inputs.

Timelapse documentation (the drawing was generated over roughly one week of machine operation, within a two-week working period, compressed into 52 seconds)

Reciprocal Latent Space is a generative drawing system that translates the urban and natural systems of Los Angeles into a slow, evolving cartography. Using GIS datasets derived from satellite sensing, the machine plots paths of infrastructure, density, and terrain onto a large wall surface, building layers of ink that accumulate like sediment.

As the drawing unfolds, moisture deliberately disrupts the marks, softening edges and introducing moments of erosion and bleed. These interruptions destabilize the precision of the data, allowing the drawing to drift between intention and accident. What emerges is a negotiation between machine logic, environmental processes, and material behavior.

The work frames the city as a living organism—constantly forming, eroding, and regenerating. By combining technological mapping with organic transformations, Reciprocal Latent Space becomes a study of urban evolution, machine agency, and the entangled systems that shape human and natural environments.

Additional process views