Topic hub
Forecasting & Time Series
Building and validating forecasts from time-ordered data: trend, seasonality, backtesting, and recognising when a model has quietly stopped working.
Forecast Accuracy Metrics: MAE, RMSE, MAPE and When Each Misleads
Compare MAE, RMSE, MAPE, and MASE for forecast accuracy, see each metric's real failure modes, and learn to backtest with rolling-origin cross-validation.
Granger Causality: A Step-by-Step Guide
Granger causality explained with a step-by-step workflow: stationarity checks, lag selection, a worked example, and nonlinear and time-varying extensions.
Holt-Winters Forecasting: When and How to Use It
A compact Holt-Winters guide for analysts: choose additive or multiplicative seasonality, initialize with 1-2 seasonal cycles, and avoid tuning mistakes.
SARIMA Model: Diagnostics and Forecasting Guide
Learn how to build a SARIMA model, choose seasonal differencing, run residual diagnostics, and forecast a monthly (m=12) time series example.
Time Series Cross Validation: A Practical Guide
Research based guide to time series cross validation. Pick reproducible splits, fix window size before tuning, and use Statohub calculators.
Time Series Stationarity: Testing & Transforms
A practical workflow for time series stationarity: plot the series, run ADF and KPSS tests, then apply the right transform before modeling.