12 FURTHER MATHEMATICS Investigating and modelling time series

12 FURTHER MATHEMATICS Investigating and modelling time series

Investigating and modelling time series Time series data are a special kind of bivariate data, where the explanatory variable is time. An example of time series data is the following table of Australian annual birth rates (average births per female) between 1931 and 2002. This dataset is rather complex, and it is hard to see any patterns just by looking at the data. However, we can start to make sense of the data by displaying it graphically. The graph we use for this purpose is called a time series plot. A time series plot is a line graph with time plotted on the horizontal axis. The variable under investigation, the response variable, is plotted on the vertical axis.

Investigating and modelling time series

Investigating and modelling time series

Investigating and modelling time series

Investigating and modelling time series Each of these trends can be explained by changing socioeconomic circumstances. Trend 1: Between 1940 and 1961 the birth rate in Australia grew quite dramatically. Those in the armed services came home from the Second World War, and the economy grew quickly. This rapid increase in the Australian birth rate during this period is known as the ‘Baby Boom’. Trend 2: From about 1962 until 1980 the birth rate declined very rapidly. Birth control methods became more effective, and women started to think more about careers. This period is sometimes referred to as the ‘Baby Bust’. Trend 3: During the 1980 s, and up until the early 2000 s, the birth rate continued to decline slowly for a complex range of social and economic reasons.

Investigating and modelling time series

Investigating and modelling time series The upward sloping trend line signals the presence of a general increasing trend. This tells us that, even though demand for accommodation has fluctuated from month to month, demand for hotel and motel accommodation has increased over time.

Investigating and modelling time series The plot reveals an abrupt change in power usage in June to July. During this period, monthly power use suddenly decreases from around 300 k. Wh per month from January to June to around 175 k Wh for the rest of the year. This is an example of structural change that can probably be explained by a change in tenants, from a family with two children to a person living alone. Structural change is also displayed in the birth rate time series plot we saw earlier. This revealed three quite distinct trends during the period 1900– 2010. These reflect significant external events (like a war) or changes in social and economic circumstances. One consequence of structural change is that we can no longer use a single mathematical model to describe the key features of a time series plot.

Investigating and modelling time series For this household, daily electricity use follows a regular pattern that, although fluctuating, averages about 10 k. Wh per day. In terms of daily power use, day 4 is a clear outlier, with less than 2 k. Wh of electricity used. A follow-up investigation found that, on this day, the house was without power for 18 hours due to a storm, so much less power was used than normal.

Investigating and modelling time series

Investigating and modelling time series

Investigating and modelling time series

Investigating and modelling time series

Investigating and modelling time series

WORK TO BE COMPLETED Exercise 6 A Odd Numbered Questions