It is well known that the situation of global climate change is alarming. The United Nation Environment Programme (UNEP) has recently reported that there will be a 2.52.9°C temperature rise above the pre-industrial levels in this century, which is far above the 1.5°C limit that would avoid the worst impact of climate change.
While the temperature in Singapore is rather constant, this does not imply that global warming has no impact on this country. The government as well as the citizens must take this threat seriously. As a data analyst, you are to construct a general forecast of temperature in Singapore based on temperature developments from the past.
In the Excel datafile “SG_Enviroment_TemperatureRainfall.xlsx”, the annual average maximum and minimum temperatures (i.e., the average of the daily highest and lowest temperature in each year) between 1960 and 2022 are recorded. You are asked to predict the temperature in Singapore for the next 10 years.
ARE YOU A STUDENT? Need Assignment Writing Help Online?
Welcome Home - Best Essay Writers and Tutors Available. No ChatGPT. No Plagiarism. CLICK HERE
The data were first collected by the National Environment Agency of Singapore. The SingStat Table Builder transformed it into the format which you can see in this Excel spreadsheet. However, in its current format, it is not suitable for time series forecasting. As a result, you should prepare the data in such way that time series methods can be applied.
Subsequently, you should explore the temperature series before carrying out any forecasting task. For this, you may generate some graphical outputs of the series data to understand temperature developments in Singapore in the last 60 years. Decomposing the series to illustrate the existence and magnitude of its components would also be a good way to assess a time series.
To forecast future temperature, you are to propose several candidate models and appraise their appropriateness for your temperature forecasting task. Please present your arguments that underpin the logic and intuition that lead you to choose these candidate models. After constructing and estimating the models, select the best model by their goodness of fit as well as their forecasting accuracy. Statistics or graphical outputs that support your decision would be helpful.
Finally, from the results you have observed in this entire analysis, discuss your thoughts about Singapore’s temperature developments in the short- and medium-term future. With what attitude should we view our future in this regard based on ONLY the forecasted result here?
Your report must not exceed 1500 words. It should contain clear elaboration of your arguments and conclusions, visualisations, and analytical results. Please provide references that you have used to support your discussions. The references may include journal articles or internet blogs.
The data preparation part is worth 10 marks. The exploration and decomposition parts are worth a total of 10 marks, and the analysis and forecasting parts 30 marks.
Your report will be assessed based on several factors: the provided insight in the discussion regarding climate change as well as the climatic situation in Singapore, the quality of writing (e.g., structure, coherence, argumentation, etc.), the quality of analysis and forecasting (e.g., the appropriate use of the model, the correct choice of the methods, the accuracy of the calculation), and conclusions.
The quality and clarity of your R codes is worth 10 marks.
Question 2
Subsequent to the analysis you have carried out in Question 1, you are to examine the relationship between the annual development of temperature and rainfall.
The idea of this analysis and forecast is based on the idea that the change in temperature of a region goes hand in hand with rainfall. Moreover, it would be interesting to see if rainfall has a lagged effect on temperature. For example, if rainfall has an immediate or delayed impact on the temperature, a forecasting model for temperature should also include historical information on rainfall as a predictor.
The Excel datafile named “SG_Enviroment_TemperatureRainfall.xlsx”, i.e., which you have used for your analysis for in Question 1, contains not only temperature data in Singapore between 1960 and 2022 but also total rainfall (in mm) of every year within the same time span. Using this dataset, you are to forecast temperature in Singapore for the next 10 years using a forecasting model that includes rainfall amount as the predictor.
For the data preparation step, append the rainfall data to the data frame that you have prepared for the tasks in Question 1 appropriately. Subsequently, study the relationship between the rainfall series and the temperature series graphically. In the next step, design an appropriate dynamic regression framework to carry out your forecasting task. Find the best transfer function and the most suitable ARIMA order for the noise term. Detailed explanation of your choice is required.
Finally, compare the forecasts resulting from the fitted dynamic regression with the forecasts from the champion model you have selected in Question 1. In terms of the goodness of fit and the forecasting accuracy, which model would you prefer here? Does the additional predictor used here increase the forecasting quality?
Your report must not exceed 1500 words and contain clear elaboration of your arguments and conclusions, visualisations, and analytical results. You must also provide references of your arguments if they have been used by other authors in other media such as journal articles or internet blogs.
The exploration on the relationship of the two series are worth a total of 10 marks, and the analysis and forecasting parts 20 marks.
Your report will be assessed based on several factors: the provided insight in the discussion regarding the relationship between temperature and rainfall, the quality of writing (e.g., structure, coherence, argumentation, etc.), the quality of analysis and forecasting (e.g., the appropriate use of the model, the correct choice of the methods, the accuracy of the calculation), and conclusions.