Correlations · Beta
What's actually making you feel like this?
Discover what actually affects your mood, with statistics computed entirely on your device.
The most interesting questions about your data require comparing trackers: does exercise lift your mood? Does meditation? Do your meds? Does eating more show up on the scale? Most apps either don't answer these, or answer them by sending your data to a server.
LogZero's Correlation module (Beta) runs seven correlations on your device: mood against exercise, meditation, your activity tags, and each medication you take, and weekly weight change against exercise level, calorie intake, and activities. Exercise and meditation are checked the same day and the next. Every result gets a strength (none / weak / moderate / strong) and a chart you can interpret yourself.
Delta analysis flags improvements and regressions over time. On Pro, the analysis window is adjustable (30, 90, or 365 days, or everything) and individual correlations can be toggled on or off.
What you get
Capabilities
Two-lag analysis
Same-day and next-day correlations. Patterns that show up the morning after.
Strength + significance
None / weak / moderate / strong, with r, sample size, and p-value on every result.
Scatter plots
See the data, not just an aggregated number. Trend line overlaid.
Delta detection
Recent improvement? Recent regression? Flag changes in the relationship.
Adjustable window
30, 90, or 365 days, or all of your history (Pro).
100% on-device math
No statistics package phones home. No data ever leaves.
In the app
A closer look
Why this is private
Cross-tracker correlation is exactly the kind of derived insight surveillance economies are built on. LogZero computes it on your phone and shows it only to you. The moat is mathematical, not marketing.
Read the privacy postureRelated
You might also use
Common questions
Common questions about correlations
How does LogZero find correlations without sending data anywhere?
The math runs locally. LogZero computes standard statistics (Pearson's r and its relatives) across seven curated correlations, on your iPhone, using only the data already on your device. No statistics package phones home. No data ever leaves.
What kinds of patterns can LogZero find?
Same-day relationships (does a workout lift your mood today?) and next-day lags (does today's exercise predict tomorrow's mood?). Mood is compared against exercise, meditation, your activity tags, and each medication; weekly weight change against exercise level, calorie intake, and activities. Each result shows its strength (none / weak / moderate / strong) with r, sample size, and p-value, plus a chart you can interpret yourself.
How much data do I need before correlations become useful?
As soon as you have about two weeks of overlapping entries between two trackers, LogZero can compute a meaningful correlation; pairs without enough data are labeled "Not enough data" instead of guessing. On Pro, the analysis window is adjustable (30, 90, or 365 days, or all of your history), so you can widen it as your history grows. The longer your history, the stronger the statistical evidence — what shows up as a weak signal in week three often firms into a moderate or strong one by month two.