Time Series Forecasting in Python – Tutorial for Beginners

An overview of time series data, defining its key components: trend, seasonality, and residuals, while explaining the importance of statistical models as a foundation for advanced forecasting.

Introduces essential baseline models like naive, historic average, window average, and seasonal naive. Explains how these provide necessary benchmarks for evaluating complex models.

Deep dive into the ARIMA (AutoRegressive Integrated Moving Average) model and its seasonal variant, SARIMA. Explains how auto ARIMA automates the selection of optimal parameters.

Covers the critical concept of time series cross-validation, emphasizing why it is superior to a simple train-test split for evaluating model performance over multiple windows.

Discusses how to include external variables (like pricing or time-based features) to improve forecasts, noting how to handle future values of these features.

Explains probabilistic forecasting using prediction intervals and discusses various evaluation metrics, including MAE, SMAPE, MASE, and CRPS, to determine model success.