Start here

What the Data Actually Says

Loading top findings…
Predictive intelligence

What Today Predicts About Tomorrow

Lagged correlations test whether today's value of one metric predicts tomorrow's value of another. These are the closest thing to causal signals in N=1 data.

Loading predictive correlations...
The Data

Data Explorer

Cross-Domain Intelligence — What Predicts What

Every metric pair tells a story. Pick two variables and see what the data says. 90-day rolling correlations, FDR-corrected, from 26 live data sources.

Metric pairs
—
Significant
—
Week
—
Why this matters
Cross-domain findings are what single-tracker apps cannot show. Your Whoop cannot tell you that your sleep predicts your glucose response. Your CGM does not know about your training load. This explorer connects 26 data sources to surface relationships that no single device can see.
Pick Two Metrics
See a pattern we missed? Submit your own finding and help build the evidence base.
Loading correlations…
See a pattern we missed?
Submit a Finding

Spotted an interesting correlation? Submit it for review. If validated, it could become a Discovery or seed a new Experiment.

Methodology

Pearson r computed over a 90-day rolling window. Pairs require a minimum of 14 matched data points. Significance testing uses Benjamini-Hochberg FDR correction at q=0.05 to control false discovery rate across all pairs tested simultaneously.

Lagged correlations test whether today's value of metric A predicts tomorrow's (or later) value of metric B. This is closer to causation than cross-sectional correlation, but still not proof. N=1 data with known confounders.

Strength labels: |r| ≥ 0.7 = strong, |r| ≥ 0.5 = moderate, |r| ≥ 0.3 = weak, below 0.3 = negligible. These are conservative thresholds appropriate for noisy biometric data.