The Whole Picture / A shared vocabulary
A little clarity
goes a long way.
Useful words for understanding your records, the connections between them, and the information behind them.
Connected wellbeing definitions
- AI persona
- An AI persona is a named communication style for an AI system. Moodwell’s coaches have different voices and generated portraits; they are not real people, healthcare professionals, or evidence of clinical review.
- Baseline
- A baseline is a defined reference period or set of observations used for comparison. It describes the available records, not an ideal state or a diagnosis. A baseline changes meaning when the logging method or the circumstances change.
- Causation
- Causation is a relationship in which changing one factor produces a change in another. Two measures moving together in a personal log does not, by itself, establish that relationship.
- Confounder
- A confounder is a factor that influences the relationship being examined and can make an association misleading. For example, a changing work schedule might affect both recorded sleep and training attendance.
- Cross-domain correlation
- Cross-domain correlation is an observed association between records from different areas of life, such as training, food, sleep, and mood. It describes how recorded measures vary together. It does not establish what caused a change, predict an individual outcome, or turn a personal tracking history into a clinical assessment.
- Data coverage
- Data coverage is the scope of the records, products, fields, or sources included in a result. A result with incomplete coverage cannot answer for the items that were excluded or unavailable.
- Event time
- Event time is when an activity or observation happened. It can differ from entry time, which is when the record was saved. Keeping the distinction matters when reviewing the order of events.
- Ingredient overlap
- Ingredient overlap is the presence of an identified ingredient in more than one recorded product. It is a description of product contents, not proof of an interaction or a judgment about a combination.
- Integration
- An integration is a supported connection that transfers or exchanges specified data between systems. An integration does not automatically provide every record type or calculate relationships between the records it transfers.
- Missing data
- Missing data is information that is unavailable for a record or period. An unlogged meal, an absent nutrient value, and a missing symptom entry remain unknown; they cannot automatically be treated as zero or as events that did not happen.
- Modeled level
- A modeled level is an estimate calculated from recorded inputs and assumptions about how a quantity changes over time. A modeled drug level is not a measured concentration and does not establish a personalized dosing decision.
- N-of-1 record
- An n-of-1 record is a collection of observations about one person. Ordinary personal tracking is not automatically a controlled n-of-1 trial; missing information, changing circumstances, and confounding still limit its interpretation.
- Normalized name
- A normalized name is a standardized representation used to connect records that refer to a defined concept. Name normalization can help match drug or ingredient records, but similarity in spelling alone does not prove that two products are equivalent.
- Observational record
- An observational record describes what was reported or measured without a controlled intervention designed to test a cause. A personal timeline can preserve a useful sequence while leaving its explanation unresolved.
- Planned versus recorded
- Planned versus recorded distinguishes an intended activity from an entry confirming what happened. A scheduled workout or dose is not evidence of completion, and a missing completion entry is not always evidence that the activity did not occur.
- Provenance in health data
- Provenance in health data is the record of where a value came from, how it was produced or changed, and which source version it represents. It helps distinguish a measurement, a label statement, a personal entry, and an estimate. Provenance makes a claim traceable; it does not guarantee correctness.
- Record and reveal
- Record and reveal is a health-tracking design principle: preserve what a person records, make relevant information understandable, and keep observations separate from medical conclusions. A timeline can document symptoms and doses without diagnosing a condition, judging a treatment, or telling someone to change what they take.
- Serving basis
- A serving basis is the quantity of food or product to which reported values apply, such as one labeled serving or a specified weight. Comparing values without matching their serving basis can produce a misleading difference.
- Source version
- A source version is the particular revision or dated release from which information was taken. Product labels, database records, and matching rules can change, so the version helps explain differences between two results.
- Time lag
- A time lag is the interval between the observations being compared. Same-day and next-day comparisons ask different questions. Selecting a lag after seeing an appealing pattern can make a chance relationship look more persuasive.
- Unchecked item
- An unchecked item is a record or product for which the relevant assessment was not completed. It cannot be treated as a completed check with no finding, and it cannot be treated as cleared.