Aglet

Triage a Weighted Average Report Measure

Weighted-average triage begins with the question the number is meant to answer. Compare per-row values, weights, and grouped outputs in a small fixture where large and small groups differ. Preserve the metric label and period filters, because an average can be mathematically valid while misleading readers about the population represented.

Establish what is happening

  1. State the averaging unit

    Write whether the report averages rows, entities, groups, periods, or already summarized values. Identify the weight field and explain whether zero or missing weights are allowed. Keep one high-weight observation beside several low-weight observations so competing results are visible.

  2. Trace numerator and denominator

    Follow the values through filtering, joins, grouping, numerator calculation, denominator selection, and display formatting. Compare weighted sum divided by total weight with the unweighted row average. Record the first stage where the report's label no longer matches the calculation.

  3. Bound affected measures

    List cards, charts, tables, exports, and comparisons using the measure. Separate surfaces with a documented weighting rule from ones that inherit a shared metric. Ask the metric owner to resolve whether the requested number is an average of observations or an average of groups.

What to carry forward

Triage is complete when the averaging unit, weight rule, first divergent calculation, affected surfaces, and metric owner are explicit. Keep the result open if the business question allows more than one average and the report provides no way to distinguish them.

Technical background: PostgreSQL documentation.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow