Data Analysis with Microsoft Excel: Updated for Office 2007

462 Statistical Methods

4 Click the Linear Trend option button to add a linear trend to the

forecasted temperature values.

5 Click the Output button and specify the worksheet Smoothed

Temperatures 2 as the output worksheet. Click the OK button twice.

Figure 11-20 shows the forecasted temperature values using two-parameter

exponential smoothing.

Figure 11-20

Forecasting

global

temperatures

with

two-parameter
exponential
smoothing

By adding the trend parameter our smoothed values have picked up the

recent trend in rising global temperatures indicated in the data. At this point

you can close the Global Temperature Analysis workbook.

Seasonality

Often time series are measured on a seasonal basis, such as monthly or

quarterly. If the data are sales of ice cream, toys, or electric power, there is

a pattern that repeats each year. Ice cream and electric power sales are high

in the summer, and toy sales are high in December.

Multiplicative Seasonality

If the sales of some of your products are seasonal, you might want to adjust

your sales for the seasonal effect, in order to compare fi gures from month

to month. To compare November and December sales, should you use the