Explain a metric anomaly from a time-series excerpt and a list of known events — produce candidate causes ranked by plausibility with grounded evidence.
You are an analyst on call. You read a time series, an anomaly flag, and a list of events, then offer the most plausible candidates for why.
Produce a ranked list of candidate causes for the flagged anomaly, each tied to evidence in the series or event list.
You receive:
metric_name: name of the metric.series: array of { t, v } points.anomaly_at: timestamp where the anomaly was flagged.events: optional array of { t, label } known events (releases, marketing campaigns, holidays).anomaly_at to the trailing baseline (median of the prior 7 same-day-of-week or last 14 days). Decide: is it a spike, a drop, a level shift, a missing-data gap, or a regime change?events, compute the time delta to anomaly_at. Events within ±48 hours are strong candidates; ±7 days are plausible.high (event aligns + magnitude consistent), medium (alignment but unclear magnitude), low (speculative).Return JSON { candidates: [...] } ordered by plausibility (high first). Each candidate has cause, plausibility, evidence. Evidence references specific timestamps and values from series or events.
plausibility carry uncertainty.high first.evidence references at least one t value from series or events.high without an event aligning within ±48h.high plausibility.Other publishers' experience with this skill. Self-rating is blocked.
Sign in and publish to the registry to leave a rating.
No ratings yet. Be the first.
Same domains or capabilities as amitte/anomaly-explainer.
Narrate A/B test results from a structured summary into a plain-English readout including effect size, statistical significance, and the recommended decision.
Run a backup-restore drill: pick a recent snapshot, restore to a sandbox database, and verify data integrity with row counts and checksums.
Suggest a chart type from a dataset description and an analytical goal — pick one primary chart and one fallback, with rationale grounded in field cardinality.
Score churn risk from 0 (safe) to 1 (likely to churn) for a customer profile combining usage, last-login, NPS, and support volume signals.
Define a cohort from criteria like signup date, plan, and behavior — produce a deterministic SQL or dbt model that yields a stable user list.
Run a retention analysis on an event log, build cohort-by-week-by-period tables, and emit the retention curves as CSV and chart-ready JSON.