BUSINESS FORECASTING

You have been employed as a consultant for a joint project by the Restaurant and Catering Industry Association of Australia and the Australian Government Treasury. As part of your role in the Business Analytics and Data Analytics team, you have been asked to forecast turnover for cafes, restaurants, and takeaway food services, as part of a wider report being commissioned by the above collaboration – on Australia’s hospitality industry.

Questions

▪ Obtain the ABS statistics for Retail Trade - 8501.0 – available at: https://www.abs.gov.au/statistics/industry/retail-and-wholesale-trade/retail-tradeaustralia/latest-release#data-download

▪ Download Table 1.

▪ For the purposes of this report you are to consider the Cafes, Restaurants, and Takeaway Food Services data. There are three series in Table 1: Original, Seasonally-adjusted, and Trend (please choose carefully throughout this report!)

▪ For the purposes of this report, only consider the data from March 2011 to February 2020 as the sample of data that is available to you – that is, ignore any recent observations.

▪ This means that the first actual observation in your Excel file is from March 2011 and your last actual observation in your Excel file is from February 2020.

▪ Use Excel and no other statistical software for the purposes of this report.

▪ You may use Minitab for constructing correlograms. This report will require two separate submissions. The numerical responses need to be submitted via a quiz tool in iLearn. The written responses need to be submitted via a PDF uploaded via Turn-It-In in iLearn. Instances of plagiarism will be dealt with according to the relevant policies and procedures. [Please turn over]4 Numerical responses to be submitted via a quiz tool on iLearn:

Exercise 1 – Application

For the purposes of this report, only consider the data from March 2011 to February 2020 as the sample of data that is available to you – that is, ignore any recent observations. This means that the first actual observation in your Excel file is from March 2011 and your last actual observation in your Excel file is from February 2020. For the Seasonally-adjusted data for Cafes, Restaurants and Takeaway Food (Series ID: A3348639K) available in Table 1: Forecast the out-of-sample values for every month in the period March 2020 – February 2021 (both months inclusive) using Holt’s Exponential Smoothing with the following parameters: alpha = 0.5 and beta = 0.2. For the seed of the level use the first observation, Y1. For the seed of the trend – take the difference of the first two observations (Y2 – Y1). Before you begin Exercise 1, let’s check that you have the right data! The average should be 3371! Once you perform Holt’s Exponential Smoothing with alpha = 0.5 and beta = 0.2, what are the following numerical values: 1. The within-sample forecast for February 2020. 2. The out-of-sample forecast for March 2020. 3. The out-of-sample forecast for February 2021. 4. The MSE. 5. The MAE. Critically think for a way to optimise alpha and beta via the MSE, and report the following values after your optimisation: 6. Alpha 7. Beta 8. The MSE 9. The out-of-sample forecast for March 2020. 10. The out-of-sample forecast for February 2021. [Please turn over]

Exercise 2 – Application

For the purposes of this report, only consider the data from March 2011 to February 2020 as the sample of data that is available to you – that is, ignore any recent observations. This means that the first actual observation in your Excel file is from March 2011 and your last actual observation in your Excel file is from February 2020. For the Original data for data for Cafes, Restaurants and Takeaway Food (Series ID: A3348636C) available in Table 1: Forecast the out-of-sample values for every month in the period March 2020 – February 2021 (both months inclusive) using Winter’s Exponential Smoothing (Multiplicative) with the following parameters: alpha = 0.5, beta = 0.2, and gamma = 0.1. For the seeds of the level, trend, and seasonal components – utilise the methods described and discussed in class. Before you begin Exercise 2, let’s check that you have the right data! The average should be 3372/3 Once you perform Winters Exponential Smoothing with alpha = 0.5, beta = 0.2, and gamma = 0.1, what are the following numerical values: 11. The seasonal component for February 2020. 12. The within-sample forecast for February 2020. 13. The out-of-sample forecast for February 2021. 14. The MSE. 15. The MAE. Critically think for a way to optimise alpha, beta, and gamma via the MSE, and report the following values after your optimisation: 16. Alpha 17. Gamma 18. The MSE 19. The within-sample forecast for February 2020. 20. The out-of-sample forecast for February 2021.

Your Exercise 3 responses should refer to Exercise 2. In addition to this, you may refer to Exercise 1. For the model in Exercise 2, given that you have the actual data for the out-of-sample period (you considered the within-sample period to end in February 2020

– but you do have data for March 2020 and onwards)

– discuss your forecasting method, your forecasts, and the business insights from these, using the following steps:

▪ Attribution

▪ Scope

▪ Application

▪ Analysis

▪ Articulation of Issues

▪ Critique

▪ Position