Averaging and Completeness

Averaging and completeness

Summary

Average / Aggregate calculates explicit statistics over timezone-aware periods. Confirm the timestamp, interval, missing markers and field statistic. Completeness is based on expected observations, not merely on the number of rows present.

Configure a job

1. Choose Average / Aggregate, select one file or a compatible batch, and inspect.
2. Open Configure averaging.
3. Confirm the timestamp/parser, source timezone, reporting-boundary timezone and input interval.
4. Review source missing-marker choices and select Marker decisions reviewed.
5. Review selected fields, statistics and output names. Confirm or override suggestions.
6. Choose Direct or Incremental, configure periods and thresholds, and review previews/completeness.
7. Select Review averaging plan, inspect full-file results, export and verify.

Direct versus Incremental

Direct calculates one target stage from source observations. Incremental runs an explicit chain of two or more stages and retains intermediate tables and completeness evidence. The two approaches need not give identical results after intermediate rejection or weighting; choose the method required by your protocol.

flowchart LR
    Input[Timestamped observations] --> Boundary[Timezone-aware period boundaries]
    Boundary --> Approach{Direct or incremental}
    Approach --> Complete[Completeness checks]
    Complete --> Strategy[Field aggregation strategies]
    Strategy --> Output[Preview / export / verification]

Loading

Periods and boundaries

Options include common/custom clock intervals, day, week, calendar month, anchored fixed 30 days, quarter, custom season, reporting year and anchored fixed year. Calendar and fixed-length periods are not interchangeable.

Set reporting timezone and relevant day/week/month/season/year starts explicitly. DST can make a local day contain 23 or 25 elapsed hourly slots. Incompatible source intervals and period chains block instead of silently dropping or reweighting data.

Completeness

The default threshold is 75%. Expected counts derive from confirmed intervals and exact period boundaries. Valid/expected counts, availability, accepted/missing state and per-stage results remain visible.

The optional 2-of-3 exception applies only to an explicitly configured Incremental stage with exactly three expected components. It is not a general alternative to the threshold. Rejected field-period results stay missing.

Statistics

Arithmetic mean is the default. Other strategies include minimum, maximum, sum, median, population standard deviation and count. Leq uses an energy average, not an arithmetic average of dB; optional positive durations enable duration weighting. Rainfall accumulation and rain-rate mean are distinct choices.

Field names may suggest Leq or rainfall behavior, but a name is not proof of unit or meaning. Review units and the applicable measurement protocol yourself.

Sources: User guide · Averaging design.