Skip to content

Guest Post: A Great Classroom Forecasting Exercise

February 19, 2013

steve harrodDr. Steven Harrod is Assistant Professor of Operations Management at the University of Dayton and can be reached at steven.harrod@udayton.edu. This is his 3rd guest post for our OM blog.

This large data set, Excel based, forecasting exercise is suitable for an hour lecture, after students have learned basic time series forecast methods in Chapter 4 of the Heizer/Render text. It gives a “real world” experience, and provides an excellent opportunity to visual the significance of error statistics in more detail. Here are instructions:

  1. Distribute the data (here is the link), but do not reveal its source.
  2. Ask the students to experiment (in Excel or Excel OM) by implementing a variety of time series forecast methods (moving average, exponential, etc.). Provide guidance, as you prefer. Give the students time to work independently (or at least in groups without you lecturing).
  3. Intermittently reveal your own progress in completing the steps on a projection screen. In a typical lecture period, you should be able to progress through two or three forecast models, and then pick one of those for error statistics.
  4. Discuss picking a “best” model according to error statistics (MAD, MSE, etc.).
  5. Pick one model, and demonstrate the calculation of the tracking signal. Chart this signal.

For discussion, ask: What are the data? Why is the tracking signal spiking? Reveal that the data are the recorded miles per gallon of a minivan at each fuel tank filling for a period of about two years. The tracking signal is spiking because the family takes vacations, and the mileage shifts from city driving to highway driving. When the tracking signal spikes, it is an indication that some fundamental change has occurred in the underlying process. The tracking signal measures whether the underlying process of the series data is stable. Since a forecast is simply the generation of a trend from a series of data points, the methodology is dependent on the underlying process being stable. If the underlying process is not stable, or experiencing a fundamental change in behavior, the forecast can not accurately predict the trend.

Advertisements
No comments yet

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Supply Chain Management Research

Andreas Wieland’s supply chain management blog for academics and managers

better operations

Thoughts on continuous improvement: from TPS to XPS

%d bloggers like this: