midTime Series
What is stationarity in time series and why is it important?
Updated May 15, 2026
Short answer
A stationary time series has constant mean, variance, and autocorrelation over time.
Deep explanation
Stationarity is a key assumption in many classical time series models like ARIMA. A stationary series has statistical properties that do not change over time. If a series is non-stationary, models struggle to learn consistent patterns. Techniques like differencing, log transforms, and detrending are used to achieve stationarity.
Real-world example
Economic indicators like GDP growth are often transformed to stationary form before modeling.
Common mistakes
- Applying ARIMA directly on non-stationary data.
Follow-up questions
- What is the ADF test?
- How do you make a series stationary?