ThinkingAIAgenticEngine·ThinkingAIAgenticEngine/scenario-skills/skills/analysis-guides/ae-user-metric-quantile/SKILL.md

ae-user-metric-quantile

Skill6 starsChanged 30 days agoNo licenceView on GitHub

Statistical methods for aggregating metric values at the user level, covering common statistics computed by "first computing a metric value per user, then running statistics over that value" (sum, average, maximum, minimum, median, quantiles, variance, standard deviation, distinct count, average per user, etc.). Trigger when users need to compute metrics such as "the 75th percentile of each user's daily usage duration", "the average of daily per-user payment amounts over the past 7 days", or "the maximum/variance/standard deviation of each user's daily consumption" — i.e., first compute a metric value per user, then run statistics over those values. The implementation chain is "metric-value tag + first/last tag + virtual event property referencing the tag (created via CLI) + Events Analysis report (published to a dashboard)", driven entirely through the AE CLI (ae-cli), not a pure SQL approach. Do not use this Skill for simple one-shot event counting without user-level aggregation, for retention/funnel/path diagnosis, or for metric alerting.