for Business Intelligence-I need you write Part b Task B (section a-c)

This report attempts to predict the number of tourists to Australia in the next 12 months. Short-term visitor arrivals data have been used to make the prediction. Time series approach is used. Smoothing techniques have been applied to smoothen the series. Winters’ multiplicative method with parameters 0.2,0.2 and 0.2 for trend, level and seasonal components are applied in the analysis. The forecasts of 12 months are made by applying this model.

This report analyses the number of tourists visiting Australia by the use of historical data available from Australian Bureau of Statistics. The purpose of this report is to find out an appropriate time series and modelling method to predict the total number of tourists visiting Australia.

This report is segregated into three sections. The first section justifies the use of short-term visitor arrivals as the time series used to predict the number of tourists in the next 12 months. The second section will introduce several smoothing approaches. The third section justifies the use of double exponential smoothing method to make predictions. The last section summarizes and concludes our findings in this preliminary report.

In this analysis, we are applying time series to study tourism growth rate in Australia, and possibly predict future forecast. The term “Time series” refers to measurements collected at a successive equal interval over certain duration of time. Time series analysis can be applied in sales forecasting, stock market analysis, census analysis, yields projection or even economic forecasting and many other phenomenon’.

Choosing short-term arrivals data from ABS is the appropriate approach to estimate the total number of tourists. It should be anticipated that the major source of tourists are tourists from abroad, given that the population in Australia remain a small percentage to the total population in the world. Estimating short-term visitor arrivals can give us the closet idea on the number of tourists, despite some tourists may visit Australia for a longer period. However, the percentage of long-term visitors should be expected to occupy a small percentage of total number of visitors to the country.

The graph represented above shows a directly proportional seasonal time series graph. As the years progressed, the number of new Australia visits is observed to be drastically increasing steadily. However, the graph has a set-back. The graph is vibrating up-and-down due to trend time series valuation of visitors in Australia. This has been brought about by different causes of visits to Australia and also confounding factors such as how long visitors stay in Australia.

The data used should be on new Australia arrivals and not people who are residence of the country. Visitors on business/school trips, leaders from different corners of the world and also immigrants should be considered as tourist (new arrivals in Australia). These populations are the most appropriate to reflect future trends of the tourism industry. It should however be noted that long-term visitors’ data is not suitable for forecast since persons on long-term visits may not necessarily be tourists in the Australia.

Tourists’ attraction spots availability

Availability of physical features, attractive and beautiful features historical aspects of a country and many more spots are likely the main component that attracts visits to a certain country.

Global Economy

Both global macroeconomic conditions and domestic economic conditions affect the number of visitors to Australia. Global macroeconomic environment recovered after the global financial crisis. Thus, private consumptions are buoyant in developed and developing countries. Australia is a favored tourist spot for tourists for its benches and natural environment. Stronger private consumption will lead to higher arrival figures. (Dharmaratne, 1995) . It should be noted that major source of tourists to Australia are from Asia and North America. Both regions experienced buoyant economic growth in the past few years. As we have observed from the time series, the growth rate(trend component) is higher when the economic growth rate of these countries were strong. Dating back to days of global financial crisis, the growth rate of tourists slowed down significant due to weaker private consumption on tourism.

Australia governance and political aspects

Governance refers to the rule of law of a country. The nature of policies available in Australia is found to be appropriate in the tourism industry. The political stability in the country is also favoring the sector and improving tourism economy.

Immigration

Immigration refers to the movement from one country to another. Favorable immigration policies in Australia has enhanced movement to-and-from the country. This has resulted to a growth in the tourism department of the Australians.

It should be noted that some of the seasonal factors should remain in the near futures. For example, tourists prefer visiting Australia during summer period (Dec- Jan) and avoid winter (Jun-Aug). It can be observed from the time series that the number of short-term arrivals dropped significantly compared to other months of the year. Economic factors such as global economic conditions and number of tourist attractions should keep the momentum of trend component continue.

In this section, three different smoothing methods is going to be used to smoothen the time series. The moving plot of the short-term visitors is shown below. The MA parameter used here is 3. It is better to capture the seasonal effect of the time series as indicated by the time series plot. (Lim and McAleer, 2001)

Another model used to smooth the series is the double exponential smoothing series. It is expected more recent values carry more weight than values in the past. It can be found that the double exponential smoothing model fits the seasonal peaks better than the moving average model.

The second model used to smoothen the series is the double exponential smoothing method. The method can capture both seasonal and trend term. The value of indicates the level constant, which is 0.87 in this optimized model. The trend constant used in this model is 0.018. It can be found that the model fits the model better than moving average approach, especially in terms of capturing the seasonal pattern.

The third method used to smoothen the series is the Winters’ multiplicative model. The seasonal, trend and level parameters used in the model is 0.2. The Winters’ multiplicative approach can smoothen the series by breaking the series into seasonal, trend and level components. This fits the time series as the series has significant seasonal component and it is expected that the seasonal component can be captured well if we need to use that model to make our forecasts.

In this section, we presented the use of Winters’ method to forecast the number of short term arrivals as well as trend models. We also compare the accuracy of other time series smoothing method discussed in the previous section in an attempt to justify the use of Winters’ method to forecast the short-term visitor arrivals, as a proxy of the number of visitors to Australia from Jul-17 to Jun-18.

When different smoothing models are evaluated, we need to identify metrics that can be used to evaluate the performance of different models. In time series forecasting, the accuracy of the model is measured by Mean Absolute Deviation (MAD) and Mean Absolute Percentage Error (MAPE).

Based on this approach, we have summarized in the table below the MAD and MAPE of models being considered in this study.

From the table above, it can be found that the Winters’ Method has the lowest MAD and MAPE. Thus, it justifies the choice of using Winters’ method to predict the number of tourists in Australia from Jul-17 to Jun-18. The forecasts are displayed in the appendix. We compare Winters’ method to Trend analysis method and found that Winters’ method can better deal with seasonal component of the time series in this case.

From the analysis above, it can be concluded that the appropriate model to make predictions on the number of tourists in Australia is the short-term visitors arrival data. However, domestic tourisms have been ignored if short-term visitor arrivals are used to predict the number of tourists in Australia. The appropriate forecasting method is Winters’ method, which allows tracking of seasonal and trend component.

Austrade, (2016) Tourism Forecasts 2016. [online] Available at: http://www.tra.gov.au/documents/forecasts/Tourism_Forecasts_2016.pdf [Accessed 16 Apr. 2017].

Australian Bureau of Statistics, (2017). 3401.0 - Overseas Arrivals and Departures, Australia, Feb 2017. [online] Abs.gov.au. Available at: http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/3401.0Feb%202017?OpenDocument [Accessed 15 Apr. 2017].

Dharmaratne, G. (1995). Forecasting tourist arrivals in Barbados. Annals of Tourism Research, 22(4), pp.804-818.

Evans, M. (2010). Practical Business Forecasting. 1st ed. Hoboken: John Wiley & Sons, Ltd.

Lim, C. and McAleer, M. (2001). Forecasting tourist arrivals. Annals of Tourism Research, 28(4), pp.965-977.

Forecasts of short-term arrivals from Jul-17 to Jun-18