Microsoft EmployeeGreat blog Christian. A tip for extending the power BI forecasting to handle growing or shrinking seasonality forecasting is to preprocess by taking a natural logarithm of historic data points and create the forecast. But to see the result just post process values with exponential function to restore values as final. Why the extra steps? Avoid having to reach for R and still handle seasonal forecasting with growth. Technically this is multiplicative rather than simply out of the box additive forecasts. Additive will only keep seasonal cycling amplitudes at constant levels.
