Tutorial Explore and Compare

Tutorial: Explore and Compare

Finding Patterns in Your Own Biometric Data

You've imported your WHOOP history (or you're wearing a strap), and now you're looking at months—or years—of metrics: recovery, HRV, strain, sleep, resting heart rate. NOOP gives you four power-user tools to interrogate that data: Explore (drill down into any single metric), Compare (overlay signals and spot connections), Insights (behavioural effects and correlations), and Trends (the longitudinal view). This tutorial walks you through real-world workflows to find patterns that matter to you.

---

Part 1: Explore — Deep-Dive Into Any Metric

When to use: You want to understand how one metric has evolved, spot outliers, and see what else correlates with it.

Step 1: Open Explore and pick a metric

1. In the sidebar, tap Explore.
2. The screen shows every signal grouped by category (Recovery, Sleep, Strain, Vitals, Activity, Body Composition).
3. Pick any metric—say, HRV (Heart Rate Variability, under Vitals). A faint dot next to the title means no recorded data; metrics with history have no marker.

What you'll see:

• The metric's full name, source (WHOOP or Apple Health), and unit.
• A chevron on the right indicating it's a link.

Step 2: Choose your window

Once you've tapped into the metric's detail dossier, you'll see a range control at the top: W (week) / M (month) / 3M (3 months) / 6M (6 months) / 1Y (year) / ALL.

1. Start with 1Y or ALL to see the big picture.
2. A caption below the control tells you how many readings you have ("N readings · ").
3. If you pick a window with no data (say, your Apple Health weight import has only 4 readings in 6 months), NOOP auto-widens to the next larger range that has data. A warning tint shows this happened: "N readings · sparse — widened to ".

Why this matters: You're always looking at real data. A short, sparse window never masquerades as "current"; instead it's honest about needing a wider lens.

Step 3: Read the hero chart

The chart dominates the centre, showing your metric over time as a line (colour-coded by category). The top-right corner displays your latest value and "as of ", so you always know the recency.

1. Hover over the chart (on macOS) to see exact values on any day.
2. The chart auto-scales vertically with padding, so the line never sits flat against the axis.
3. A colour gradient reflects the metric's category: recovery rings blue-green, strain glows red-orange, HRV is purple.

Tip: If your data is sparse (e.g., weekly weight readings), the window expands gracefully, and the chart still reads clearly because NOOP zooms relative to the latest data point, not "today".

Step 4: Scan the stat tiles

Below the chart is a row of uniform tiles showing:

• Average — mean over the window, with a sparkline and the count of days.
• Min — lowest value recorded.
• Max — highest value recorded.
• Latest — the most recent reading and its date.
• Δ vs prev — how this window compares to the previous equal-length window. Tinted green if the change is "good" (e.g., recovery ↑), red if "bad" (recovery ↓). If there's no prior window to compare (e.g., you're looking at all-time history), this tile says "all history" instead.

Real example: If your HRV average over the last month is 45 ms and the previous month was 42 ms, the Δ tile shows "+3 ms" in green (HRV up is good).

Step 5: Check "What correlates"

At the bottom is the Pearson correlation scan: NOOP's cross-catalog engine automatically compares your chosen metric against every other metric in the system over the visible window. It surfaces the top 6 strongest relationships (|r| ≥ 0.30, requiring at least 10 overlapping days).

Each row shows:

• The correlated metric's name, category, and sample count.
• An r-bar — a horizontal bar showing the strength and direction. Positive (r > 0) fills to the right in green; negative (r < 0) fills to the left in red; negligible correlations sit pale and narrow.
• The exact r value (e.g., "+0.67").

Reading the bars:

• Very full bar, green → strong positive: when your metric rises, the other tends to rise too.
• Very full bar, red → strong negative: when your metric rises, the other tends to fall.
• Barely visible bar, pale → weak or negligible: they move independently.

Example workflow: You notice your HRV and recovery are the two strongest correlations (r ≈ +0.85). That makes sense—HRV is a key driver of NOOP's recovery score. But you also spot that resting HR correlates negatively (r ≈ −0.60): higher RHR tracks with lower HRV. That's a useful insight: when your RHR creeps up, your HRV likely drops, and recovery may follow.

Tip: Nothing under |r| ≥ 0.30? The card says "Nothing in the catalog moves clearly with over this window. Widen the range to surface relationships." Try a longer window.

---

Part 2: Compare — Overlay and Correlate

When to use: You want to see two or more metrics on the same timeline, spot how they move together, and get a plain-English read on their relationship.

Step 1: Open Compare and select metrics

1. In the sidebar, tap Compare.
2. You'll see an "Overlay 2–4 signals" section with a + Add metric button.
3. Tap the button—a grouped menu appears with every category (Recovery, Sleep, Vitals, Activity, Body Composition).
4. Pick your first metric, say Recovery. It appears as a removable chip with its colour swatch.
5. Pick a second metric, say Sleep Performance. You can now pick up to 2 more (max 4 total).

What the chips show:

• Each metric's name and a coloured dot (stable categorical colours: mint, cyan, purple, amber).
• A small × to remove it.

Step 2: Choose the time window

A range control sits beside the "Add metric" button: W / M / 3M / 6M / 1Y / ALL (defaults to 1Y).

• Pick 1Y to see a full year of correlation.
• A caption shows "N readings across metrics · ".
• If any selected metric is sparse in that window, the caption warns "· sparse widened".

Step 3: Study the normalized overlay chart

NOOP plots all selected metrics on one chart, min–max normalized to 0–1 so different units share a y-axis. Each series is a distinct colour.

1. Hover over the chart (macOS) or tap (mobile) to see exact values:
   • A vertical crosshair marks the day you're looking at.
   • Dots appear on each series showing where it falls.
   • A floating tooltip lists every metric's real (non-normalized) value for that day.

Example: You're comparing Recovery (0–100), Weight (150–180 lbs), and Sleep Performance (40–95%). On one day:

• Recovery is 72 (normalized to ~0.6: middle-upper area).
• Weight is 165 (normalized to ~0.5: middle).
• Sleep is 88 (normalized to ~0.9: near the top).

The chart shows all three lines at these heights, and the tooltip reads: "Recovery 72 · Weight 165 · Sleep 88%".

Legend below the chart: Each metric shows its real min–max range (e.g., "Recovery: 45–89 · Sleep: 32–96%"), so you always know the true spread.

Step 4: Read "How They Move Together"

Below the chart, every pair of selected metrics gets its own correlation card. With 3 metrics you'll see 3 cards (Recovery ↔ Sleep, Recovery ↔ Weight, Sleep ↔ Weight); with 4 you'll see 6.

Each card shows:

• A pair header with both metric names and their coloured dots.
• r = ±value (e.g., "r = +0.34") in a tint matching the correlation strength.
• A plain-English sentence interpreting the relationship.
• A footer with the sample count and strength/direction summary.

Example cards:

Recovery ↔ Sleep Performance:

r = +0.67 (strong positive) over 345 shared days.
When sleep performance rises, recovery tends to rise — a strong positive link.

Weight ↔ Recovery:

r = −0.28 (weak negative) over 203 shared days.
When weight rises, recovery tends to fall — a weak negative link.
No clear relationship — they move largely independently.

Strength words used:

• r < 0.1: negligible (no clear relationship).
• 0.1 to 0.3: weak.
• 0.3 to 0.5: moderate.
• 0.5 to 0.7: strong.
• ≥ 0.7: very strong.

Direction words:

• r > 0: positive ("tend to rise together").
• r < 0: negative ("one tends to fall when the other rises").

Step 5: Act on what you see

Use this to answer real questions:

• "Does my weight drop when my recovery improves?" Compare them side-by-side.
• "How much does strain yesterday carry into tomorrow's recovery?" (See Insights for lagged correlations.)
• "Does my sleep efficiency track with my HRV?" Overlay them and read the r value.

Tip: Weak or negligible correlations are also meaningful. If you expected a strong link and it's weak, that's data: those two metrics are more independent than you thought.

---

Part 3: Insights — Behaviour Effects and Metric Relationships

When to use: You want to understand (a) which logged behaviours move your outcomes, or (b) how key metrics relate at the system level.

Step 1: Open Insights

1. In the sidebar, tap Insights.
2. The screen has two halves: Behaviour Effects (top) and Metric Relationships (bottom).

Behaviour Effects (requires journal)

If you've imported a WHOOP export that includes journal entries (Alcohol, Caffeine, Late meal, Meditation, etc.), this section lights up.

1. Pick an outcome metric using the segmented control: Recovery / HRV / Sleep / RHR.
2. NOOP splits your days into two groups:
   • Days you logged that behaviour ("With").
   • Days you didn't ("Without").
3. It then compares your chosen outcome between the two groups.

What each effect card shows:

• Behaviour name (e.g., "Alcohol") with a tinted circle (green = positive effect on your outcome, red = negative).
• A significance pill: SIGNIFICANT (p < 0.05 and n ≥ 5 days in both groups) or EXPLORATORY (interesting but not yet robust).
• Plain-English sentence: "On days you logged 'Alcohol', Recovery was 12% lower (avg 61 vs 69, n=140 vs 498)."
• Two StatTiles: "With" and "Without" showing the mean values, sample counts, and % change (Δ).
• Effect size (Cohen's d): A small-to-large magnitude word. d = 0.2 is small; d = 0.8 is large.

Real example:

• Meditation: "On days you logged 'Meditation', Sleep performance was 8% higher (avg 87 vs 79, n=52 vs 398)." Significant, green, d = 0.41 (moderate).
• Late meal: "On days you logged 'Late meal', Sleep performance was 6% lower (avg 74 vs 79, n=89 vs 398)." Exploratory (not yet significant), d = 0.28 (small).

Interpretation tips:

• Significant effects have passed a statistical gate and deserve your attention.
• Exploratory effects are hints; keep logging if you want to reach significance.
• The effect-size word (negligible / small / moderate / large) tells you magnitude, independent of significance.

Metric Relationships (no import required)

A curated set of four Pearson correlations computed from your stored series:

1. Sleep performance ↔ Recovery — how a good night tracks with next-morning recovery.
2. HRV ↔ Recovery — the engine behind NOOP's recovery score.
3. Resting HR ↔ Recovery — expected to be negative (high RHR, lower recovery).
4. Recovery → Next-day recovery (1-day lag) — how much one day's recovery carries into tomorrow, reflecting strain carry-over.

Each row shows:

• Metric pair (e.g., "Sleep performance ↔ Recovery").
• r value (e.g., "r = +0.67") in a colour-coded swatch (green = positive, red = negative, yellow = weak).
• Significance pill: "p < 0.05" (significant) or "n.s." (not significant).
• Correlation bar — a centred bar: zero in the middle, fills left (negative) or right (positive) by magnitude. Hover to see the exact r.
• Plain-English reading: "A strong positive relationship (r = 0.67, n = 87 days)."
• Blurb explaining what the pairing probes.

Why this matters: These relationships are the "foundation" of NOOP's analytics. If HRV ↔ Recovery is r > 0.7, you know HRV is a strong driver—track it. If Recovery → Next-day recovery is weak, strain may not carry forward as much as you'd expect.

---

Part 4: Trends — The Long View

When to use: You want to see how your metrics have evolved over weeks, months, or years; spot seasonal patterns; and understand your personal baseline.

Step 1: Open Trends and pick a range

1. In the sidebar, tap Trends.
2. A range control sits at the top: W / M / 3M / 6M / 1Y / ALL (defaults to 3M).
3. A caption shows "Trailing 90 days" (or "All history" for ALL).

Step 2: Review the hero recovery chart

The first large chart shows Recovery over the selected range, with:

• A line tracing your recovery score day-by-day.
• A filled area under the line (blue-green gradient).
• Footer stats: Avg, Peak, Low, Days recorded.

Interpretation:

• A rising line = improving recovery.
• A sawtooth pattern = normal variability; spikes up after good nights, dips after hard training.
• A sustained dip = fatigue accumulating or life stress.

Step 3: Scan the small multiples

Three Daily signals cards sit below:

Heart Rate Variability (ms):

• Fallback range: 20–120 ms.
• Footer shows mean / min / max.
• Trends upward = parasympathetic tone improving.

Resting Heart Rate (bpm):

• Fallback range: 40–80 bpm.
• Opposite direction matters: when RHR drops, recovery typically improves (lower baseline = better).
• A creeping upward trend often signals illness or excessive stress.

Day Strain (/ 21):

• Range: 0–21.
• Spikes = high-load training or stressful days.
• Sustained high strain (15+) without recovery days signals overtraining.

Each card auto-scales to show your recorded data, so a metric with tight clustering still renders clearly.

Step 4: Study the recovery year heat-strip

At the bottom is a calendar heatmap of recovery scores:

• Each small square is one day.
• Colour reflects recovery (red = depleted, gold = baseline, green = peaked).
• Shows either the past year (default) or all history (when range = ALL and you have 2+ years).
• Hover (macOS) to see exact recovery scores by date.

Use this to spot:

• Seasonal patterns. Did summer always see better recovery? Winter worse?
• Anomalies. A block of red squares might indicate an illness or period of overtraining.
• Trends. Is your overall pattern drifting greener (improving) or redder (declining)?

---

Part 5: Practical Workflows

Workflow A: "I felt great last week. What changed?"

1. Open Explore → Recovery.
2. Set to W (week). Notice the spike.
3. Open Compare, add Recovery + Sleep Performance + Resting HR.
4. Set to W. Hover over last week—see if sleep and RHR moved favourably.
5. Open Insights → behaviour effects, pick Recovery. Did you log anything (meditation, good sleep, lower alcohol) those days?

Workflow B: "My recovery has been low for a month. What's correlating?"

1. Open Explore → Recovery → 1Y.
2. Scroll to "What correlates". Spot negative correlations (things that drop when recovery drops).
3. Example: Strain (r = −0.68) and Resting HR (r = −0.52) both strongly negative.
4. Open Trends → 1Y. Watch the recovery dip and strain spike simultaneously.
5. Open Insights → behaviour effects. Did you log stressors or changes (caffeine, late meals, less sleep)?

Workflow C: "Does my training load carry over?"

1. Open Insights → Metric Relationships.
2. Look at Recovery → Next-day recovery (1-day lag). If r is high (e.g., 0.6+), yesterday's recovery predicts today's—strain carries over.
3. Open Trends → 1M. Visually trace strain spikes and see if recovery dips the next day.
4. Open Compare → add Day Strain + Recovery. Hover over high-strain days; is next-day recovery visibly lower?

---

Disclaimers & Honest Notes

• Not medical advice. Correlations and behaviour effects are patterns in your data, not clinical assessments. Do not use them to diagnose illness or change medical treatment.
• Approximations. NOOP's recovery, strain, and sleep stages are approximations of published methods, not reproductions of WHOOP's proprietary models. See ANALYTICS for details.
• Sparse data is OK. NOOP auto-widens sparse windows so you always see real data. If a metric has only a handful of readings, the visualizations gracefully expand to include them.
• Correlation ≠ causation. A strong r between two metrics means they move together, not that one causes the other. Example: recovery and HRV are strongly correlated because HRV is a component of NOOP's recovery score, not because one drives the other independently.
• Significance thresholds. Behaviour effects require p < 0.05 and n ≥ 5 in both groups (with/without). This is a safety gate to avoid spurious "significant" findings from tiny samples.

---

Quick Reference: Colours and Patterns

Element	Meaning
Tint in range caption	Warning (amber) = window auto-widened due to sparseness.
Correlation bar	Green (right) = positive; red (left) = negative; full = strong; narrow = weak.
Behaviour effect tint	Green circle = positive effect; red circle = negative effect. Bright = significant; muted = exploratory.
Recovery heat-strip	Red = depleted (~0–34%); gold = baseline (~35–66%); green = peaked (≥ 67%).
Chart line colour	Reflects metric category: recovery = blue-green gradient, strain = red-orange, HRV = purple, etc.

---

Cross-References

• Features — comprehensive feature list.
• ANALYTICS — transparent math behind recovery, strain, HRV, sleep, and correlation.
• FAQ — troubleshooting and common questions.
• Strap Support and Pairing — if you're pairing a live strap.

---

NOOP is independent, offline, and not affiliated with WHOOP. Your data stays on your machine. Happy exploring.