Step 1
Orient the scope.
Start on the public overview. It tells you which layer you are looking at and what kind of claim it is allowed to carry.
Open overviewViews
This page goes straight to the dashboards already materialized in the repo. The landing is the entry point; these are the analytical surfaces underneath it.
Reading path
Orient the scope first, read the validation boundary second, and only then open the live layer that matches the question you are asking.
Step 1
Start on the public overview. It tells you which layer you are looking at and what kind of claim it is allowed to carry.
Open overviewStep 2
Before opening a dashboard, read the validation and evidence pages so signal, transport, and limitations stay explicit.
Open validationStep 3
Only then go to the analytical dashboard that matches the layer you need: night, event, daily, or non-stress.
Open layer indexSurface index
The index below keeps the analytical surfaces legible before you drop into the full dashboard view.
Dashboards
These cards preserve each layer’s own status, best visible signal, and route to the underlying dashboard.
night
Nightly routes are the core operational layer. They carry held-out signal, interval calibration, governance, and deployment status.
| heldout_r2 | 0.4350 |
| heldout_mae | 1.7972 |
| heldout_rmse | 2.2791 |
| coverage_pct | 80.2118 |
| interval_width | 5.1123 |
| transport_r2 | -0.1106 |
event
Event models remain separate from nightly interpretation. They are portable for acute physiological stress, not for next-day sleep claims.
| mae | 0.6775 |
| rmse | 0.8182 |
| r2 | 0.3164 |
daily
Daily challenge cohorts show where temporal memory helps within cohort, even when transport still fails across cohorts.
| mae | 1.3178 |
| rmse | 1.6673 |
| r2 | 0.1685 |
| rows | 20112 |
| participants | 925 |
| feature_count | 91 |
| best_model | extra_trees |
nonstress_daily
This layer keeps anxiety, depression, and somatic support visible without overstating clinical portability.
| mae | 1.2722 |
| rmse | 1.6985 |
| r2 | 0.3454 |
| rows | 4818 |
| participants | 196 |
| feature_count | 292 |
| best_model | hist_gbm |
nonstress_family
Portable non-stress families remain research references. Anxiety is closest to usable; depression and somatic burden are still below transport readiness.
| portable_mean_r2 | 0.0202 |
| route_replay_r2 | -0.0466 |
| coverage_pct | 79.8680 |
| rows | 303 |
| participants | 103 |
| feature_count | 10 |
Direct access
Use these cards when the layer is already clear and you want to move straight into the analytical surface.
Live layer
Open the dedicated dashboard for this layer without passing through the broader public narrative.
Open dashboardLive layer
Open the dedicated dashboard for this layer without passing through the broader public narrative.
Open dashboardLive layer
Open the dedicated dashboard for this layer without passing through the broader public narrative.
Open dashboardLive layer
Open the dedicated dashboard for this layer without passing through the broader public narrative.
Open dashboardLive layer
Open the dedicated dashboard for this layer without passing through the broader public narrative.
Open dashboardParticipant-facing access
This is not a separate consumer product. It is a weekly longitudinal rendering layer built on top of approved nightly stress routes and the same structured report engine used elsewhere in the stack.