Time Series

A time series is a sequence of a metric's values recorded in time order; the very thing that lets you see a number change across days and weeks.

A time series is the basis of every trend analysis: without time order a metric is just a lone number; with it, a story. The key difference from cross-sectional data is that the observations are not independent — today's value depends on yesterday's — and that dependence is both its power and its trap. Three components hide inside a time series and must be told apart: the long-term trend, the repeating seasonal pattern, and random noise. If you do not separate these three, you mistake an ordinary weekend dip for a real problem. The time axis's scale is a decision, not a detail: the same data looks volatile at an hourly view and smooth at a weekly one, and choosing the wrong scale can hide or fake a pattern.