📈 Windows (in Time-Series)

Windows (in Time-Series)#

Fixed spans over which time-series features or aggregates are computed.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

In time-series analysis and monitoring, a window is a slice of time over which you compute a statistic — a mean, sum, baseline or anomaly score. The window defines how much past data counts, and it can stay fixed, slide forward, or expand.

The three types#

A fixed window always covers a set period — “weekly sales, Monday to Sunday”. A rolling (or sliding) window moves forward step by step, recomputing as it goes — “a 7-day rolling average of temperature” — and is the workhorse for smoothing, baselines and anomaly detection. An expanding window starts at a point and grows as data arrives — “cumulative average sales since launch”.

Baseline vs current#

Monitoring usually compares two windows. A baseline window — a rolling average over the last N weeks — defines what “normal” looks like. A short current window — the last 24 hours to 7 days — shows whether recent behaviour has deviated from that baseline.

An example#

For API monitoring, a rolling 8-week baseline gives an average latency of 200ms. The current 24-hour window shows 350ms. Because 350ms far exceeds the baseline, the comparison flags performance degradation — the same windowed-comparison logic that underlies drift detection and leading indicators.

Theme: Signal Processing & Time Series · All terminology

Hint

Mind map — connected ideas

Sliding Window (Rolling Window) Cross-Validation · Expanding Window Cross-Validation · Time Series · Monitoring Pipelines · Drift Detection · Leading Indicators

Hint

More in Signal Processing & Time Series

ARIMA (AutoRegressive Integrated Moving Average) · Bayesian Time Series · Forecast Error · Forecasting Benchmarks · Forecasting Competitions · Log-Space · Low-pass Filtering · LSTM — Long Short-Term Memory Networks · M-Competitions (Makridakis Competitions) · Naïve Baseline Forecast · Prophet — Time Series Forecasting by Facebook (Meta) · Seasonal Lag · Seasonality · Signal Processing