ORION SYNAPSE

Data visualisation defaults to charts because charts are safe. But a public atrium is not a dashboard — nobody stands still to read an axis. The piece had to be immediately legible as beauty and only gradually legible as information.

Bind the data to physics instead of geometry. Volume becomes density, latency becomes drift, and bursts become pulses that travel along the constellation. Nothing is labelled; the structure teaches itself through repetition.

Standing in front of the piece, visitors first read weather, then rhythm, then meaning. Because the source is live and unrepeatable, regulars notice the difference between a quiet morning and a peak hour without ever being told what they are looking at.

Deep field, thin light, no chrome. The composition holds an enormous amount of dark so that a single bright pulse carries real weight. Generative typography is used sparingly, as texture and timestamp, never as a caption.

• Streaming ingest pipeline with windowed aggregation
• GPU particle systems with instanced link rendering
• Generative typographic layer for temporal markers
• Long-running stability harness for unattended operation

• Data audit — deciding what is honest to show and what is noise
• Mapping studies from metric to physical behaviour
• Endurance testing for continuous multi-month playback
• Graceful degradation when the feed stalls or drops

The mapping architecture became a reference for our later data work: physical behaviour first, encoding second. It is the clearest demonstration that an information system can be composed rather than merely plotted.

• Data art installation
• Streaming data pipelines
• GPU particle systems
• Visual experience design