nonstationary

Imagine you are studying a stream of data over time. At first it stays in a narrow range, then it starts to drift, jump, or vary more wildly than before. That is the kind of situation where "nonstationary" is most often used. The word usually means that something does not remain statistically stable over time.

It appears most often in technical fields such as statistics, economics, engineering, and data science. A nonstationary process is one whose key properties change over time. The average may drift upward or downward. The amount of variation may grow or shrink. Repeating patterns may weaken, strengthen, or move. A radio signal with changing noise levels, a financial market series with long-term trends, or climate data with shifting patterns can all be described this way.

This matters because many models work best when the data behaves in a steady way. If the data is nonstationary, a model may give weak or misleading results unless that change is handled directly. For example, economists often test whether inflation, prices, or output follow a stable pattern or keep shifting over time.

Outside technical writing, people sometimes use the word more loosely for something that is not fixed or steady. Still, its clearest and most common meaning is about change over time in data, signals, or systems rather than simple physical motion. So when you see "nonstationary," think less of an object moving through space and more of a pattern that refuses to stay the same.