Homework answers / question archive /   Semester Project           Due 12/3/2014   The objective of the project is to simulate a problem you might face in business—developing the best 2 year strategic plan forecast for company sales

  Semester Project           Due 12/3/2014   The objective of the project is to simulate a problem you might face in business—developing the best 2 year strategic plan forecast for company sales

Economics

Share With

---

 

Semester Project

          Due 12/3/2014

 

The objective of the project is to simulate a problem you might face in business—developing the best 2 year strategic plan forecast for company sales.  Forecasting is an important process in almost all businesses and is an important skill to possess as an employee.  There are no project examples so the creation of the report and your organization of the material will be part of your grade.  Executives like a report that is easy to read and reference so I suggest that material from Minitab required to support your observations or points appear in your discussion.  The project is due by midnight Central Time on 12/3/2013. You are strongly encouraged to spread the work throughout the semester.

 

1. Selecting independent (X) variables for your project[1]

Hopefully you have all passed this stage.  You need to pick a minimum of three X variable quarterly data sets of a least 50 observations each. Make sure the data are NOT seasonally adjusted.   The data relationships should be logical and not spurious.  For example, do not use the price of hamburger to forecast auto sales.  (See the Proposal instructions for more detail).

 

2. Writing an executive summary

After you pick your topic, it is a good idea to write a brief abstract to help focus your thoughts. The executive summary should not be more than a page. It should include

 

1. 1. State your forecasting problem- develop the best forecast for the company revenue (Y variable).
   2. Information about and description of your data
   3. General description of your proposed models
   4. Findings –The best forecast method

 

3. Body of the paper

The paper will contain detailed explanations of your thought processes and methodology.

 

After the executive summary the report should begin with an introduction that is essentially your proposal.  The problem statement should be presented early in the introduction.  Your problem is to develop the best two year forecast for your company revenue.  You should mention that residual analysis and error measures will be used to determine the best model to use.  Next you may want to mention the four forecast methods that you will explore.

 

Since one of the methods (multiple regression analysis) uses other variables to forecast company sales you should next provide the results of your proposal and analysis of your variable selection hypotheses. 

 

After the introduction the body of your paper will include all four major forecasting techniques.  I suggest you may want to have a section for each of the four models you will present.  In each section you should include the model description and a time series plot of the model fit period residuals, ACFs for the fit period residuals, histogram of the fit period residuals and error measures for the fit period.   You should include the forecast appended to the fit period data and a plot of the forecast and hold out along with forecast period error measures.  You may also wish to provide a time series plot of the residuals to note error characteristics as well. These elements are part of the extra credit assignments.

 

The four forecast methods are:

 

1.        Smoothing for the Y variable (Chapters 4): You need explain which smoothing model type was selected and why. 
2.       Decomposition for the Y variable (Chapter 5): You need to explain the cycle factor adjustment.
3. Box-Jenkins for the Y variable (Chapter 9): You need to explain what you did to make the data series stationary.

 

4. Multiple Regression forecast of the sales variable (Y) (Chapters 6, 7 and 8): You will use the independent variables (Xs) to help explain the variation in the sales variable of interest. You need to start looking for those explanatory (X) variables!! (See the Project Proposal for more detail)

 

Generally, at every stage you need to answer the following questions for each forecast method:

 

• What did you do (ex: Winter’s method with these parameter values…)
• Why did you do it? (e.g. The data series indicated seasonality as shown by time series plots and autocorrelation functions.)
• What did you find? (e.g. Interpretation of your results for fit and forecast of residual analysis and error measures)
• What is your conclusion? (e.g. The model produced an acceptable or non-acceptable forecast)

Remember, this is a business project report for your company – Do Not Show failed model attempts.  Show only the best model for each variable for each forecast method.  We will investigate four forecast methods so you will have to submit results for only four models.

 

Finally, after the discussion of each model you should have a conclusion section that presents at table of error measures for each model for the fit and forecast periods.  Point out the model that produces the lowest forecast period error in the conclusion. 

 

 

 

4. Appendix

Your appendix should contain all the relevant supporting data and table output from the models or other analysis.  It is better to divide the appendix into parts where each part is representing the output from each methodology (ex: Appendix A: X variable Data and citations, Appendix B other variables used, Appendix C: Sales Revenue forecast outputs).  Don’t forget to name or number the appendices, you will need to refer to them when you are talking about a specific methodology.

 

5. Final Report

Your final report should include an executive summary of your findings which will include the conclusions (A brief two paragraph statement of which of the techniques works best and why for the assigned company revenue variable?). You need to have an introduction to talk about your topic in greater length and X variable selection.  It will include a section for each of the four model types and a conclusion.  Do not forget the appendix with relevant data and citations.  Be detailed and complete in all your sections.  This is a case of more is better than less as you explain to me what you have done.