Throughout the term you will progressively complete all three case studies and the instructor will randomly draw one of your group’s three submissions to grade for your final project. Upon submission you will supply a PDF or Word report that specifically addresses each of the case study assignment questions along with the following areas:
| Section | Standard | Possible Points |
|---|---|---|
| Define Goal |
Discuss the overarching goal of the case study and how forecasting helps to address the problem. Specifically, discuss: 1.1 The descriptive versus predictive goals. 1.2 The forecasting horizon and updating needs. 1.3 How the forecast will, or could be, used. 1.4 Required automation. |
5 |
| Exploratory Data Analysis |
Provide a background of the data to include: 2.1 Temporal frequency & granularity. 2.2 Missing data, outliers, unequally spaced series, and other abnormalities. 2.3 Presence of time series components such as level, trend, and seasonality. 2.4 Be sure to provide sufficient visualizations to illustrate these features. |
10 |
| Pre-Process Data |
Once you've identified all the attributes of your data discuss: 3.1 How you corrected for missing values, outliers, unequally spaced series, and other abnormalities. 3.2 If you aggregated your data to a different level than provided. 3.3 Any other data pre-processing steps you performed. |
5 |
| Partition Data |
Outline and defend your approach for partitioning your data into training and validation sets to include: 4.1 How you decided on the length of the training and validation sets. 4.2 Whether you performed cross-validation. |
5 |
| Apply Forecasting Methods |
Discuss your methodical approach to applying your forecasting methods: 5.1 Did you apply a simplistic, basic approach to act as a benchmark? 5.2 What attributes in your data make your selected forecasting methods appropriate? 5.3 How and why did you tune any parameters? 5.4 Provide both visual and numeric outputs of your forecasting models. |
10 |
| Evaluate & Compare Performance |
Illustrate the performance of your model(s) by visualizing and discussing: 6.1 Training and validation residuals. 6.2 Training and validation accuracy measures. 6.3 Confindence intervals. 6.4 Based on your evaluation and performance assessment, which final model did you select? 6.5 Provide the mathematical notation for your final model. 6.6 Use the final selected model to forecast the required k steps ahead (provide both visual and numeric outputs. |
10 |
| Summary |
7.1 Summarize the forecasting problem you addressed. 7.2 Summarize how you addressed this problem statement (the data used and the methodology employed). 7.3 Summarize the interesting insights that your analysis provided. 7.4 Summarize the policy or decision-making implications as a result of your analysis. 7.5 Discuss the limitations of your analysis and how you, or someone else, could improve upon it. |
5 |
| Formatting & Other Requirements |
8.1 Analysis is systematic - complicated problem broken down into sub-problems that are individually much simpler. Analysis is efficient, correct, and minimal. 8.2 All case study questions in the back of the book were answered. 8.3 Achievement, mastery, cleverness, creativity: Tools and techniques from the course are applied very competently and, perhaps,somewhat creatively. Perhaps student has gone beyond what was expected and required, e.g., extraordinary effort, additional tools not addressed by this course, unusually sophisticated application of tools from course. 8.4 Analysis is reproducible meaning the instructor can fully re-create the analysis in the report due to the level of details provided or because the students provide the analysis code (.Rmd file) as an appendix. |
10 |
Total possible points: 60
Due no later than: Wednesday, December 13, 11:59PM ET