Lab / Datasets
The lab

The data behind the smell fingerprints.

Our first dataset comes from the analysis pipeline that defines the lab's data contracts. Real bench, chamber, and field recordings are being collected now.

Pipeline dataset

The training ground for the electronic-nose pipeline — built to lock down the data contracts, feature shapes, and model bar the hardware has to meet.

90
runs
10,800
rows
5
events
15s
sample period
Eventsclean · coffee · iqos · perfume · vape
Sensorsbme688 · sgp40 · sht40 · mics6814 · pms5003
Splitscenario-heldout · 60 train / 30 test
Features114 per sample

Bench & field

Collecting
Hardware datasets — in progress

Real bench, chamber, and field smell recordings are being gathered under the AER protocols. Dataset cards and access will appear here as runs come in.

Placeholder — data currently being collected.

Access & references

For dataset access or collaboration, contact [email protected].

  • 01phase1_artifacts/phase1_gate_report.json — dataset summary & metrics.
  • 02phase1_artifacts/run_bundles/ — sample run bundle (manifest + samples).
  • 03GitHub — XoAnonXo/aeralyte.