Forecasting with Indices
Geist Industries manufactures various types of snow shovels for distribution to several states in the United States, mostly for small retail businesses. The corporate headquarters and manufacturing plant are located in Anchorage, Alaska with distribution plants in 21 states. The company has recently hired a new CEO who has asked that I make a seasonal demand forecast based on the last four years of winter historical inventory data that has been recorded. The winter historical inventory data is based on actual demand in numbers of snow shovels manufactured by Geist Industries. To minimize the data and make it more manageable for analysis, I have devised a twelve month average for each of the four years and plotted the resulting data to forecast the upcoming year’s demand in Figure A. The following formula was used to calculate the twelve month average individually for years one through four:

Total Demand (Actual) for Year (January thru December) = Individual Year Average 12

(Figure A) I have extended the trend line to forecast the 5th year demand, in this case to provide a forecasted average that follows a similar trend to that which has been displayed in the past four years. The trend line that has been established displays an R2 value of 0.9553 allowing a conclusion that the data set is relatively useful in forecasting future demand. Using the 12 month average for the last four years of data we can safely assume that next years average demand will be potentially somewhere in the neighborhood of 49,000.
Analyzing the Historical Inventory Data Figure B illustrates the Winter Historical Inventory Data. To make a determination regarding how to best construct a forecast and index the data must first be reviewed for any inaccuracies or data that may need to