metrics-analyzer

Provide the CSV file path or paste the data directly. The analysis loads the data with pandas, computes summary statistics, detects trends and anomalies, breaks metrics down by segment, and produces ranked actionable insights in a saved Markdown report.

Anomalies are detected by flagging data points more than 2 standard deviations from the mean, plus sudden drops, spikes, or flatlines. Each anomaly includes the observed versus expected value and a hypothesis for the cause, labeled as speculative when appropriate.

Yes. If you ask a specific question, the analysis answers that question first, then provides the broader trend, anomaly, and segment analysis. Questions can be phrased in plain English.

Datasets under 30 data points are flagged as too small for meaningful analysis. Messy data with missing values or inconsistent formats is cleaned before analysis, and the cleaning steps taken are reported back to you.

No charts are generated. Instead, the output describes in words what a chart would show, so you can visualize the data later in your preferred tool. The full analysis is saved as a metrics-analysis-[date].md file.