Live config loaded — values shown in teal reflect what is currently deployed on this server.
AGS Documentation
Back to Dashboard

Data Lifecycle src/models

Consent, collection, processing, storage, export, exclusion, withdrawal, and retention in one research-data path.

1. Before consent
No telemetry may be written. The participant can choose language and read orientation, but study interaction rows are blocked until consent.
2. Enrollment data
Phone is converted to an HMAC token. Demographic data is coarse: age group, literacy, prior web use, language, and self-efficacy.
3. Session data
Session records hold condition, template, device/capture mode, timestamps, status, exclusion state, and completion/abandonment outcome.
4. Event data
Interaction events capture timing, errors, help, idle, navigation, and device-appropriate typing/touch signals. Idempotency uses client event IDs.
5. Derived data
Signal normalisation, readiness, governance, adaptation, and outcome classifiers produce auditable derived rows.
6. Research operations
Exclusions, withdrawals, exports, config changes, content changes, and team actions write audit entries.
7. Export and analysis
Exports are English-only, pseudonymous, and audit logged. Excluded sessions remain in raw exports but should not enter primary analysis unless specified.

Data classes

ClassExamplesHandling
IdentifierPhone HMAC, study ID, session UUIDNo raw phone in database; use study/session IDs in analysis.
Participant attributesAge group, language, literacy, prior web useCoarse and research-relevant only.
Behavioral signalsClicks, pauses, errors, help, typing/touch metricsValidated by schema and separated by device type.
Derived research dataReadiness, governance rule, adaptation outcomeComputed server-side and retained for reproducibility.
Operational dataAudit log, export log, alert resolutionUsed to defend study integrity.