[pic]
Liquor Sales Forecast
[pic]
MGSM960 Information and Decision Analysis
John Rodwell
North Ryde, Term 2 2006
Prepared by:
Warren Creighton 30335949 creightonw@dairyfarmers.com.au
Paul Sedgman 40770109 paul.sedgman@defence.gov.au
Table of contents
1 Executive summary 3
2 Industry overview 4
3 Issue identified 4
4 Model objective 4
5 Data selection, source and expectations 5
5.1 Dependent variable 5
5.2 Independent variables 5
5.2.1 Direct variables 5
5.2.2 Indirect variables 8
5.3 Demographic factors 9
5.4 Additional variables tested and omitted 10
5.5 Additional variables considered 12
5.6 Correlation matrix 13
5.7 Test models 14
6 Forecast model 14
6.1 Summary output 15
6.1.1 Adjusted r-squared 15
6.1.2 MAD 15
6.1.3 MAPE 15
6.1.4 Zero intercept 15
6.1.5 Coefficients 16
6.1.6 P-value 16
6.2 Interpretation of model 16
6.3 Investigation of significant variances 17
6.4 Additional testing 18
7 Limitations of the model 18
8 Conclusion 18
9 References 19
Appendix A '' Source data summary statistics 20
Appendix B '' Graphical representation of forecast results 21
Appendix C '' Test model: Salary and wage earners 22
Appendix D '' Test models 23
Executive summary
Across a large national retail chain, profits and other financial indicators are often difficult to compare. The aim of this quantitative analysis was to determine a model, which could identify under or over performing stores in relation to each other, separate from market conditions. Management armed with this information could use it to:
• Assess new store designs or locations,
• Identify the factors resulting in above average performance in order to emulate their performance in other stores,
• Identify under performing stores to investigate whether the performance is due to internal or external factors.
The data used was from a...