Module 30

Time series and forecasting

Advanced module (after graduation): dates, patterns, lazy forecasts, lag features, splitting by time

When you finish this module you can say:

I can spot the trend and the weekly cycle in a daily series, build the lazy forecasts, train a model on lag features, measure it without breaking the time order and forecast next week.

By the end of this module you will be able to build lag features from a time series, measure your model without breaking the time order and compare your forecast with the lazy line.

Lessons

6 lessons · ≈ 64 min

1. A date is not a number10 min
2. Seeing the pattern10 min
3. Lazy forecasts10 min
4. Lag features11 min
5. Splitting by time11 min
6. Project: next week12 min · project

Practice exercises

Short writing exercises, independent of the lessons: Go to the exercises

1. No shipping on weekends
2. How many days were far off?
3. The lazy forecast for next week