How did you handle seasonality and external influences in time series?

To handle seasonality and external influences in time series, the following approaches are usually applied:

• Decomposition of the time series — splitting the series into trend, seasonality, and residuals (for example, using the STL method or classical additive/multiplicative models).

• Using moving averages for smoothing and identifying seasonal patterns.

• Models with seasonal components — for example, SARIMA, which directly account for seasonality.

• Including external regressors (exogenous variables) — if external factors are known (holidays, weather, marketing campaigns), they can be added to the model to improve forecasts.

• Analysis and adjustment of outliers, which may be caused by external influences.

Example: if analyzing sales, one can identify seasonality by months, consider holidays as external variables, and build a SARIMAX model for forecasting.