analyze-my-data

What problem does it solve?

It turns unclear business performance and data issues into rigorous, evidence-based conclusions with explicit ship/kill/iterate guidance for experiments.

Core Features & Use Cases

• Experiment readouts with lift and decisions: Computes significance, confidence intervals, guardrail deltas, and returns a clear call (ship/kill/iterate/inconclusive-extend) rather than vague impressions, e.g., evaluating whether an A/B test result is strong enough to act on.
• Anomaly sweeps against rolling baselines: Detects metrics deviating past per-metric thresholds (with baseline shown) and proposes likely causes tied to recent operational context.
• Data-quality (DQ) audits for reliability: Checks null rates, duplicates, freshness expectations, and referential integrity to flag which tables or joins are likely breaking downstream analysis.

Quick Start

Use analyze-my-data to generate an experiment readout for an A/B test by providing your warehouse query (or pasted variant aggregates), the hypothesis, primary metric, and guardrails, and ask for a ship/kill/iterate/inconclusive decision.