FORECASTINGStart at the route map, or jump to any stop
Layout

Forecasting,
one idea at a time.

A journey from “what is a forecast?” to the models people actually use.

Each stop covers one concept. You read the idea, play with a diagram until it clicks, take away one line, and check yourself with a question.

How the forecasting journey works. Illustrative Northstar examples, not measured business results.
Part of the journeyWhat it meansNorthstar example
One running example

Use one business story to connect methods to real planning questions.

Northstar forecasts daily sales with weekly shopping patterns, underlying growth, and a Black Friday peak whose date moves each year.

Simulated teaching data
  • Constructed data helps explain which ingredients created the chart.
  • It is not a fitted business forecasting service.

Northstar's simulated components are known, so the diagrams can reveal trend, seasons, and random variation separately.

Build the ideas in order

Later stops build on earlier distinctions, while the route map still lets you revisit any topic.

Understand Northstar's uncertainty and baselines before comparing ETS, ARIMA, and more complex seasonal pipelines.

31 stops · 0 marked as understood

Take the first step
31 CONNECTED STOPS

The route map.

Follow the line, or choose any stop.

Forecasting journey route mapPART I · FOUNDATIONSPART II · MEASURINGPART III · SIGNALPART IV · MODELSPART V · MORE MODELSPART VI · IN PRACTICE01A forecast is a range02Trend, seasonality &noise03Trend or seasonality?04Decomposition05Baselines06Picking a growth rate07Holdout & backtesting08Error metrics09Autocorrelation10White noise &residuals11Stationarity &differencing12Base effects13Simple exponentialsmoothing14Holt: adding trend15Holt–Winters:seasonality16Fourier terms17Events & holidays18Model families19ETS: the framework20ARIMA & autoregression21SARIMA & SARIMAX22Dynamic harmonicregression23Prophet & changepoints24MSTL + ETS25Prophet vs DHR vs MSTL+ ETS26Prediction intervals27Drivers & leakage28Hierarchical forecasts29Machine learning &combinations30Forecast, plan &target31A forecasting workflow
Go deeper with FPP3.

Original ecommerce examples and teaching diagrams, informed by Hyndman & Athanasopoulos’s Forecasting: Principles and Practice (3rd ed.). Each stop links to further reading. These are learning tools, not production model fits.