Time-Based Data

About this unit

Much business and economic data is recorded over time: monthly sales, yearly production, daily temperatures, quarterly profits. Such data, arranged in order of time, is called a time series. This unit explains why values in a time series change and splits the change into four components: the long-term secular trend, regular seasonal variation, longer cyclical swings and irregular, unpredictable movements. You learn how to analyse a single-variable time series, how the components combine in additive and multiplicative models, and how to measure the trend using the freehand, semi-average, moving-average and least-squares methods. Measuring the trend allows simple forecasting, which is the main practical purpose of the unit.

How to prepare

• Learn the four components with one everyday example for each; many questions describe a situation and ask which component it shows.
• Memorise the additive (Y = T + S + C + I) and multiplicative (Y = T × S × C × I) models.
• Practise computing 3-year and 5-year moving averages and fitting a straight-line trend by least squares with coded years.
• Know the merits and limits of each trend method.