Notes on data sonification. In my daily job, I am often watching over multiple streams of logs and metrics. Processing all this information just by eyes sounds like a waste of resources. We've built data visualisation tools so that we can utilize our visual cortex for "free" data pre-processing. Why not use other senses?

Thinking about a rich data sonification system. Brain dump:

• For raw textual logs, flow rate as a sub-bass signal. More logs, higher frequency or a filter opening up
• Alternative for the bass, an ambient sample, something from Brian Eno's Music for Installations, rising in volume with the logs flow rate
• A log filtering engine, as simple as mapping "regex → sound". All matches are fed into a sound generator as MIDI-like events, with some debouncing applied
• Automatic determination of a "nominal" flow rate of all logs or just filtered events. Second-order derivative and some inertia? Or maybe just a simple moving average?