ASSESSMENT 2 GUIDE

COM Ml190

Data, Insight s and Decisions

Term 1, 2025

Assessment 2: Predictive Analytics Project

Context of assessment tasks

Following the initial data exploration in Assessment 1, the senior management of Dolphin Theme Park has approved further analysis of activities and sales using an expanded data set. These data have now been extended, with extra variables being added. The core task will involve a comparison of predictive models and subsequent recommendations on how to use and improve these to inform. future sales.

You and your team have been tasked with investigating alternative algorithms for predicting visitor revenue. The General Manager (GM) has given you the authority to use your expert judgment to make the necessary modelling choices but has outlined an overarching research plan for you and your group to follow:

•     Currently, the GM has assigned a regression model for each of your team members to run in Stage 1 with the current focus on predicting revenues.

•    As a team, you need to meet and compare the results of each of your regression models performed in Stage 1 and perform. a quick comparison of all your team members’ models.

•     In Stage 2, you can formulate a more reliable model using a regression tree or multiple linear regression (MLR) to predict the revenue; in other words, you can improve on the suggested models provided by the GM in Stage 1.

•    Your evaluation of predictive performance should include a comparison of the preferred predictor with that generated by your chosen regression-based predictor.

•     Based on this analysis, make recommendations on whether the use of such algorithms in initiatives targeting visitors with the ultimate objective of generating more revenue.

Stage 1:Task Instructions

While reporting is done in two stages students are encouraged to commence their collaboration within their group early in the process before the submission of the Stage 1 individual reports.

Complete the  2-page  individual task  on the template  provided  in  Appendix  B  using  a  data  set  specific  to  your Assessment 2 project group. This first-stage submission contains key inputs into the group work that will result in the single group report produced in Stage 2. Students who do not submit a complete, legitimate attempt at this assessment will not be awarded marks for Stage 2. [i.e.: legitimate attempt means that the Appendix B + R Script has been duly uploaded on Turnitin within the prescribed due date]

This individual task in Stage 1 will be separately assessed. However, the marks and feedback for Stage 1 will be released together with Stage 2.

The Appendix B outlines the three key areas of the assessment 2-stage 1:

•     you are required to perform. data wrangling and cleaning:

•     split the sample into training and test data sets

•     run the assigned model based on training data

•     evaluate the model's predictive accuracy under both training and test data.

In Stage 1, a new personalised dataset has been generated with the ‘ Pass’ variable being removed and new variables added-namely ‘ Revenue’, ‘ Rain’, and ‘Promotion’, as defined in the revised data dictionary in Appendix A. Revenue aggregates the revenue from rides and shows, merchandising, and food and beverages. Rain is the categorical variable whether there is rain or not. Promotion is the binary variable that indicates whether the promotion campaign was targeted or not before the visitors entered the park. The GM has suggested a base regression model for the prediction:

Revenue = βo  + β1Agei  + β2passTypei  + β3 Genderi  + μi

Each member of the respective team shall use one of the following modified models, as long as two members do not use the same model:

A: Add ‘HostRating’ to the base model.

B: add ‘QueueTime’ to the base model.

C: add ‘ FoodBeverage’ to the base model.

D: add ‘Merchandising’ to the base model.

If there are two members in a team, they should choose A and B. A group of three members shall choose A, B, and C.

Stage2: Task Instructions

As a group, use the results from Stage 1 to produce a report to the executive management. You will use R to explore a dataset that includes the same expanded personalised dataset as in Assessment with extra variables (see attached Appendix A Data Dictionary). A group-specific data set will be determined by nominating the ZID of one of the group members to generate the single data set used by all group members in both stages of the project.  Details for obtaining the personalized group-specific data set will be provided on Moodle.

Your Stage 2 group mark will be common to all students in your group who have submitted a complete, legitimate attempt at Stage 1.

The group task is outlined below:

•     Present a brief comparison of the Stage 1 results of your team members by comparing the accuracy of the Stage 1 models, to inform. the choice of data and preferred predictive model to use in predicting revenue.

•     You are expected to generate a more robust model predicting revenue.

•     You are required to use data visualisation in your report.

•    Your report should explain your strategy, subsequent analysis, and the recommendations that follow.

•     Include a critical evaluation of the predictors' strengths and weaknesses and any potential improvements, as management will continue to focus on this. Critical evaluation can be conducted by comparing the accuracy metrics of alternative models on both test and training data sets.

•     Use the Assessment 2 marking rubric for the Stage 2 group submission as a tool to check your work before submission and to ensure that you have addressed the assessment task in full.

Your Stage 2 submission should be a report to the General Manager. Use the initial report provided in Assignment 1 to guide you. Your Stage 2 Report should be approximately the same length (i.e.: 4 pages) and structure (i.e.: introduction, body, and recommendation) as that report, although the section headings of the body will likely change, and the emphasis on graphical presentations is likely to be less. Ultimately decisions about presentation are to be made by your group.

Your group must also submit a separate file containing the R code used to conduct your analysis and generate your visualisations. Your submission will be deemed incomplete and given a mark of zero if such a file is not included.

Overall, the assessment is designed to help you develop your skills in using R for data analysis and in communicating insights from such exercises.