How Daylogue Generates Insights

Daylogue surfaces patterns in your emotional life. If you are going to trust those patterns, you should understand how they are generated. This page walks through the full methodology: what data is captured, how patterns are identified, how narratives are written, and what Daylogue does and does not claim about accuracy.

What is captured per check-in

Each check-in can capture the following data points:

• Mood (1-10 scale). How you feel overall. This is the one metric that is always captured.
• Energy (1-10 scale, or null). How much physical and mental energy you have. This is optional. If you do not explicitly set it or mention it, it is recorded as null, not as a default value.
• Stress (1-10 scale, or null). How much stress you are experiencing. Same optional handling as energy.
• Qualitative text or voice input. What you actually said during your check-in. This is the richest data point and what the AI uses for follow-up questions, extraction, and narrative synthesis.
• Photos (optional). Images you attach to a check-in to capture a moment.
• Timestamps. When you checked in. Used for time-of-day and day-of-week pattern analysis.
• Focus Area tags. Topics that matter to you (relationships, work, health, etc.) that help contextualize your check-ins.

How patterns are identified

Daylogue uses two approaches to find patterns: statistical aggregation across your numeric data, and AI synthesis across your qualitative entries.

Statistical aggregation

Your mood, energy, and stress scores are aggregated across multiple time windows: 7-day, 30-day, and 90-day. Within those windows, Daylogue calculates:

• Day-of-week analysis. Are Mondays consistently harder? Does your energy peak on certain days? This analysis reveals recurring weekly rhythms.
• Metric correlations. How sleep relates to mood, how stress relates to energy, how one part of your life affects another. These correlations are surfaced as observations, never as causal claims.
• Tag co-occurrence. Which Focus Areas tend to appear together, and how different tags correlate with your metrics. If work-tagged check-ins consistently show higher stress, that is worth noticing.

Null values are filtered out before any aggregation. If you did not report energy on a given day, that day is excluded from energy calculations rather than being counted as a zero or a default. This is a deliberate data quality decision that prevents phantom data from distorting your patterns.

AI synthesis

Beyond numeric aggregation, the AI reads across your qualitative entries to identify themes that numbers alone would miss. Recurring topics, shifts in how you talk about certain areas of your life, emotional undercurrents that persist across check-ins. This synthesis powers the narrative feature and the deeper pattern observations.

How narratives are generated

Daylogue's narrative feature reads across multiple check-ins and synthesizes them into a written reflection about your recent days. This is not a summary of one entry. It is a synthesis of patterns across time.

Narratives are written in a reflective and warm tone. They notice themes, highlight shifts, and connect threads you might not have seen on your own. The AI is instructed to observe, not to judge, and to use language that invites reflection rather than prescribing action.

Narratives are serialized. Each one builds on the ones before it, so your story deepens over time. The first week's narrative has less to work with than the third month's. This is expected. Patterns become more visible with more data.

Data quality

Pattern quality depends on data quality. Daylogue takes several steps to ensure the data underlying your patterns is honest:

• Null over default. If you do not provide an energy or stress value, it is stored as null, not as a midpoint default. Early versions of Daylogue defaulted to 5 when values were not provided, which created phantom data in aggregations. This was identified and remediated. Historical entries affected by this issue are flagged with data quality markers.
• Confidence levels. Each extracted data point carries a confidence level: mentioned (you said it explicitly), inferred (the AI interpreted it from context), or not captured (no signal was present). This metadata ensures that patterns are built on what you actually said, not on assumptions.
• Null-aware aggregation. All averages and trend calculations filter out null values before computing. A week where you checked in three times out of seven calculates averages from those three check-ins, not from seven entries where four are zeros.

What Daylogue does not claim

Being transparent about what Daylogue does means being equally transparent about what it does not do:

• No claim of clinical accuracy. Daylogue is a wellness tool. Its pattern detection is observational, not validated against clinical benchmarks. Patterns are useful for self-reflection, not for making medical decisions.
• Correlation, not causation. When Daylogue observes that your mood is lower on days with less sleep, it is surfacing a correlation. It is not claiming that poor sleep caused your low mood. There could be a third factor driving both. Correlations are presented for your reflection, not as explanations.
• Self-reported data has limits. All of Daylogue's data comes from what you tell it. If you consistently underreport stress or overreport energy, your patterns will reflect that. Daylogue works best when you are honest with it, and it does not attempt to correct for self-reporting bias.

The ethics of insight generation

How you frame an insight matters as much as the insight itself. Daylogue's AI prompts are designed to notice, not to label. The language is deliberately non-clinical:

This is not about being vague. It is about being precise in a way that respects the boundary between self-awareness and clinical language. A wellness tool that uses clinical terminology risks leading people to conclusions they are not qualified to draw and that the tool is not qualified to support.

Daylogue surfaces what it sees. You decide what it means.