Insights is the analysis layer. It takes every spin you logged and everything the app collected around it, and answers three questions with statistics rather than hunches: what happened before, what worked after, and what is already going right on your spin-free days.
That air-pressure finding came from models that treated each patient as their own baseline rather than comparing one person with another.
Gürkov et al., PLoS One 2016For each trigger, symptom and metric the engine compares the days before your spins with your normal days using Fisher's exact test, tests a few different lags, and corrects for the number of tests. Where it can compare, a row reads much more likely, more likely, slightly more likely, or no clear link, with three confidence dots for the strength of the signal. Where there are not enough normal days to compare yet, it says seen before and shows the count. Every row always carries the honest number: in 6 of 14 spins, 43 percent. There is no invented multiplier.
Some things only bite together, and the Patterns page lists the pairs that raised your risk more than either did alone. On the other side, What Worked scores every medicine and relief on the relief you reported and how often you used it, shows which ones you used together, and the conditions each works best under. Guided exercises land here too, because finishing one is logged as a relief on the spin.
What's Helping is the positive side of your data: the habits and steady signals that show up on your spin-free days, and what drives them. What changed compares the last 30 days with the 30 before across your spins, your logged items and your daily habits. And Monitor is where the graphs live, one card per thing you track, each opening its history with your spins marked.
Insights. What Happened ranks each factor with how much more likely a spin was and in what share of them, the dangerous combinations underneath, then What Worked, What's Helping, What changed, What Were You Doing and How Did It Impact You. Monitor holds one history graph per thing you track.
About five logged spins before the engine has anything to say; until then the empty states tell you to keep logging. What it reports is correlation, stated as correlation: whether a spin was more likely when a factor was present, and in what share of spins, with the judgement left to you and your clinician.
The app says it needs about five logged spins before it has anything to say. The empty states tell you to keep logging until the patterns show, and the Accuracy card needs at least seven days of gauge data.
Correlation, stated honestly. The app tells you whether a spin was more likely when a factor was present and in what share of spins, and leaves the judgement to you and your doctor.
In Monitor, one card per thing. Insights holds the analysis: What Happened, What Worked, What's Helping, What changed, What Were You Doing, How Did It Impact You, Accuracy, Recommendations and the Full Report.
