Open data for sleep science
AirwayLab is building the largest open dataset of anonymised PAP therapy metrics. Every data point is contributed voluntarily, stripped of identifiers, and stored in the EU. We believe better data leads to better therapy — and we want to work with researchers who share that goal.
What's in the dataset
Glasgow Index
- Overall score (0–8)
- 9 component scores: skew, spike, flat top, top heavy, multi-peak, no pause, inspiratory rate, multi-breath, variable amplitude
- Per-session and duration-weighted nightly averages
WAT (Wobble Analysis Tool)
- FL Score: inspiratory flatness percentage
- Regularity: Sample Entropy on minute ventilation
- Periodicity Index: FFT power in 0.01–0.03 Hz band
NED Analysis
- NED mean, Flatness Index, Tpeak/Ti ratio
- M-shape detection percentage
- RERA events per hour, Estimated Arousal Index
- Combined FL percentage, H1/H2 split comparisons
Oximetry (optional)
- ODI-3% and ODI-4% (desaturation indices)
- Heart rate surge counts at multiple thresholds
- Coupled cardio-respiratory events
- Desaturation time, SpO₂ summary statistics
All metrics are computed from ResMed SD card flow waveform data. Raw waveforms are available for a subset of contributors who explicitly opted in to waveform sharing. Machine settings metadata (device model, pressure settings, mode) is included where available.
How data is collected
Browser-only analysis
A PAP user uploads their SD card data. All analysis runs in their browser — nothing is sent to a server.
Voluntary opt-in
After seeing their results, the user can choose to contribute anonymised metrics. This is always opt-in, never automatic.
Anonymisation
Identifying information is stripped: no dates, no names, no device serial numbers. Only aggregate breathing metrics and machine settings are stored.
Secure EU storage
Data is stored in an EU-hosted database with row-level security. The anonymisation code is open source and auditable.
How to collaborate
We're looking for researchers and clinicians who want to use real-world PAP therapy data to advance sleep science. Here's what we can offer:
- Access to anonymised, aggregate breathing metrics from thousands of real-world therapy nights
- Validation partnerships for our analysis engines (Glasgow Index, WAT, NED) against clinical gold-standard measurements
- Joint research on flow limitation patterns, RERA detection, and the relationship between breathing metrics and patient outcomes
- Open-source codebase (GPL-3.0) — adapt our algorithms for your research, contribute improvements back
Data privacy and ethics
AirwayLab handles health data. We take that seriously.
- GDPR-compliant. All data stored in the EU (Supabase, eu-west region).
- Every contribution requires explicit, informed consent. No silent collection, no pre-checked boxes.
- The anonymisation pipeline is open source — audit it yourself.
- Users can request deletion of their contributed data at any time.
Get in touch
Interested in collaborating? We'd like to hear from you.
Or email us directly at [email protected]
Medical Disclaimer
AirwayLab is not a medical device. It is provided for educational and informational purposes only. Always consult qualified healthcare providers regarding sleep therapy.