
Time series forecasting cheat sheet
A free two-page, explained cheat sheet (PDF) for time series forecasting. It uses one cafe sales example with charts to teach trend, season, stationarity, baselines, ETS, SARIMAX, Prophet, machine learning, backtesting and scoring. No sign-up.
Download PDFWhat is on it
- What a time series is, with trend, season and noise shown on a real example
- Stationarity explained with a before and after chart
- Baselines to beat, including seasonal naive
- A table to pick the right model for your situation
- Short code for exponential smoothing, SARIMAX, Prophet and machine learning
- Backtesting drawn as a picture, with TimeSeriesSplit code
- A worked MAE, RMSE, MAPE and MASE example with real numbers
- Mistakes that fool you and a 7-step recipe to follow every time
Good to know: Always compare against a seasonal naive forecast. If a complex model cannot beat it on a time-based backtest, it is not worth shipping.
Learn it step by step: Take the free time series course, eight short lessons with code and practice projects.
Questions
Is this time series cheat sheet free?
Yes. Download the PDF with no sign-up and no email needed.
Who is it for?
Students and beginners. Every idea has a plain-words explanation and an example from one cafe, so you can follow it without a statistics background. Basic Python helps for the code.
Which libraries does it use?
pandas, statsmodels, scikit-learn, Prophet and statsforecast. Install the ones you need with pip.
Can I print it?
Yes. It is made for two A4 pages.