Question

Month	Sales (2005)	Sales (2006)
Jan	5	21
Feb	8	20
Mar	10	29
Apr	18	32
May	26	44
Jun	35	58
Jul	28	46
Aug	20	32
Sep	14	27
Oct	8	13
Nov	6	11
Dec	26	52

a.Plot the data using Minitab. What kind of pattern do you observe?

b.Develop a trend line equation for this data (Use Minitab).

c.Decompose the above time series using a multiplicative model. Store the trend line, detrended data, seasonals, seasonally adjusted data. Then, calculate CI (Hint: CI = Deseasonalized data / Trend), C as a centered 3-moving average of CI values, and finally I (=CI/C).

d.Next, compute forecasts for the next 6 months using the trend line equation and the seasonal indices you have computed. (Hint: What should you assume for C and I components?)

Answer 1

Original data

(a)

The nature of the plot suggests that there are effects of both tend and seasonality in the data.

(b)

(c)

Go to Stat --> Time Series --> Decomposition

Output

Calculate CI

Output

Find C and I

(d)

Period	C	I	T	Forecast
25	1.301089	1.035986	37.59	50.66792
26	0.598973	0.986695	38.634	22.83281
27	0.942295	1.152262	39.678	43.08118
28	1.218979	0.788972	40.722	39.16396
29	1.415261	0.503309	41.766	29.75048
30	1.40553	0.993587	42.81	59.78487