Forecasting over time: from sales to energy
A time series is a set of measurements recorded in sequence over time.
What is it?
Time series are measurements recorded in sequence over time: temperatures, sales, electricity use or traffic. Models use past patterns to estimate what is likely to happen, taking account of seasons, trends and sudden changes.
An everyday example
A grid operator forecasts how much electricity we will use tomorrow evening and adjusts generation to match.
Why does it matter?
Good forecasts cut costs and risks in energy, stock and transport. A bad forecast can lead to shortages or waste.
A common misconception
That a forecast is information from the future. It is an estimate that can fail when the world changes in a way the past data never showed.
In more detail
What can be forecast?
Energy demand, traffic, stock levels or other quantities that change over time.
Why is time special?
Past values, seasons, trends and sudden events can all affect future values.
Is a forecast certain?
No. It makes sense to show a range of possibilities and to track errors over time.
How do we test the model?
We test it on later periods that it did not use during training.
Remember: A forecast is an estimate of the future, not data from the future.
Reliable sources
Last reviewed 1 November 2026. An explainer, not professional, legal or medical advice. For rules and figures that change, check the official version.