Sanity-Check a Metric Definition and Find Double-Counting Risks
ENData & Analyse
Stress-test a metric definition to expose ambiguity, double-counting, and edge cases before it goes into a dashboard.
data-qualitydouble-countingkpimetrics
LA@lacauze
0
Aide-nous à améliorer Prompédia
On mesure l'usage du site de façon 100% anonyme (aucune donnée personnelle, jamais revendue) pour l'améliorer — pour les visiteurs avec et sans compte. Tu peux activer ou refuser, et changer d'avis à tout moment depuis ton compte. En savoir plus
20 prompts
Stress-test a metric definition to expose ambiguity, double-counting, and edge cases before it goes into a dashboard.
Design a robust parsing strategy that extracts structured fields from inconsistent log lines without losing data.
Compare two datasets that should match, quantify every discrepancy, and explain the likely cause of each gap.
Turn a vague data request into a complete analysis brief that surfaces blind spots before any work starts.
Translate a business question into a well-structured pivot table with the right rows, columns, values, and filters.
Break a time series into trend, seasonality, and noise, then produce a forecast with honest uncertainty.
Generate a complete, readable data dictionary that documents every column, its meaning, type, and constraints.
Choose a defensible outlier-detection method for your variable and qualify whether each anomaly is an error or a signal.
Transform raw numbers and chart outputs into a clear, persuasive narrative for a non-technical audience.
Build a cohort retention analysis with the right metrics and a plain-language reading of what the numbers actually mean.
Design a reproducible, leakage-free feature pipeline for your ML task, with transforms fit only on training data.
Translate a business goal into a focused KPI tree and an actionable dashboard layout that drives decisions, not vanity metrics.