Project & Methodology

A personal entry connects a voluntarily selected perception with severity, time, approximate location, an expected weather phenomenon, and a prediction window. Free-text notes remain private.

Categories create a consistent journal structure without asserting a health meaning. People describe their own perception; the system does not assess the person.

1. The user records a perception and a weather expectation.
2. The selected time window passes without the result being edited retrospectively.
3. Hourly Open-Meteo values are retrieved for the rounded location and time window.
4. Defined thresholds for precipitation, pressure, temperature, humidity, wind, and storm indicators produce a match, partial match, or no match.

The result only describes whether the expected weather phenomenon was measurable during the selected window. It does not establish a medical or causal relationship.

The map presents aggregated signals within a selected radius. Public signals are spatially rounded and combined. Exact locations, identities, and personal notes do not appear on the map.

Rank levels refer to how earlier weather expectations compared with later weather data. They are a playful quality indicator, not an assessment of a person's health or credibility.

Many voluntary, spatially rounded reports can make collective knowledge visible: Which perceptions and expected weather phenomena are mentioned more often in a region at the same time? The app combines this shared experience as anonymized, aggregated trends.

This collective intelligence is an additional perception signal, not an alternative or official weather forecast. Reports do not come from a representative observation network, and their volume, distribution, and accuracy may vary greatly. Suitable weather services and official warnings remain essential for weather, travel, and safety decisions.

One entry is a snapshot. The journal becomes more useful when observations are recorded over several weeks with the same routine. That means using a reasonably consistent severity scale, choosing a clear prediction window, and recording the weather expectation before that window has passed.

1. Record the perception and severity as soon as practical.
2. Select only the weather phenomenon that is genuinely expected.
3. Choose a realistic window such as 6, 12, or 24 hours.
4. Do not revise the prediction later to match what happened.
5. Interpret several comparable entries together, not one entry alone.

This does not turn the journal into a health measurement. It simply separates a documented expectation from later memory and makes recurring personal observations easier to review consistently.

At 08:00, a person records a perceived pressure in the head with severity 5 and expects a notable air-pressure change within 12 hours. After the window ends, the system retrieves available hourly values for the rounded location. Reaching the defined threshold produces a match; a change close to that threshold can produce a partial result.

The result does not identify what caused the perception. It only shows whether the earlier weather expectation and the later available weather values align under the same rules. Sleep, stress, nutrition, activity, and many other influences are not controlled by this comparison.

• Predictions are stored with their creation time and a fixed time window.
• Evaluation follows documented thresholds rather than a free-form judgement.
• Missing weather values are not automatically treated as a failed prediction.
• Public map points are rounded and presented as grouped values.
• Test signals are demonstration data and are not presented as real user reports.

Even a long series of matches does not prove causation. It may provide a personal question for continued self-observation. Health concerns belong with an appropriately qualified healthcare professional, not with this journal.

Useful questions include: Does the same perception appear more often before a particular weather change? Does the match rate differ with a shorter time window? Are there months with more or fewer entries? These questions concern the user's own record and should always be read alongside the number of observations.

The journal cannot support conclusions about other people, diagnoses from ranking levels, or claims that a weather condition must cause a particular perception. The community map is an approximate overview, not a representative population study.

The current weather view and retrospective comparison use weather data from Open-Meteo. Map presentation uses OpenStreetMap tiles.

Weather models, observation networks, rounded coordinates, and subjective entries can be incomplete or imprecise. Results are therefore journal prompts, not scientific or medical conclusions.

• As few required fields as practical.
• Approximate rather than exact public location display.
• No publication of personal notes or identities.
• Account deletion and transparent privacy information in the app.
• Advertising remains disabled until policy review and a Google-certified consent solution are complete. Any later ad placement is intended only for public information pages, not the personal journal or community signals.