1 BISM2204 Methods of Business Analytics Project Plan – Briefing Notes Background The course BISM2204 Methods of Business Analytics has three assessment items: a project plan, a project report, and a final examination. These notes outline my expectations for the project plan or proposal (hereafter, the “project plan”). Before doing so, it is useful to briefly review the domains and methods of business analytics and highlight broad expectations of the project report (more details on the report later in the Semester). Briefly recall the business analytics matrix: application areas or domains of business analytics in one dimension and methods of business analytics in the other. Domains include accounting/financial analytics, people/talent analytics, operations analytics, marketing analytics, social media analytics, and supply chain analytics, etc. The methods of business analytics include data visualisation, predictive analytics, prescriptive analytics, text analytics, deep learning, etc. This course focuses on the methods of business analytics, and the project plan/report has a particular focus on predictive analytics. The course BISM2201 Principles of Business Analytics – the companion course for BISM2204 – focuses on application areas for business analytics, among other topics. In effect, the project plan and report ask you to make a case for an investment in a business analytics project and to conduce the project – at least to the level of a credible “pilot project.” The focus is on regression techniques for predictive analytics (predictive analytics is a broad field that includes both regression techniques and machine learning techniques – the course BISM2205 Predictive Analytics covers these methods in detail with a focus on machine learning for business). Some example applications of predictive analytics include the following. (1) Estimating the fair value for the remuneration of directors of Australian public companies given company performance. (2) Establishing the Google ratings consumers give to fast service restaurants given their qualitative comments. (3) Predicting the future value of a company’s share price given the company’s cash flows and discretionary expenditure on marketing, and research and development. These are but two of the many possible directions a project on predictive analytics (using regression techniques might take). A Context for Study This Semester, you are encouraged to focus on Spotify as the hypothetical client for project plans/project reports. I will leave you to do some background research (“desk research” or Google searches) on Spotify and the music industry. One of the key challenges for music streaming services (and other streaming services such as Netflix) is building engagement with their subscribers. A proxy for engagement is the extent to which subscribers enjoy or like the songs to which they listen. We will not have access to Spotify’s song library or its metadata, but we do have access to the Million Song Dataset (MSD). “Song hotness” is one of the key variables in the MSD), and we can use this variable as the focus output for our regression models. I will separately introduce you to the MSD, and challenge you to think about the appropriate feature variables. (You can choose a context other the music industry for your project plans/project reports; however, you have to convince me there is a dataset relevant to the question you wish to address and you can access that dataset, etc.). Think of the plan/report as a pilot project for the senior leadership team at Spotify. Its aim should be provide “proof of concept” for your ideas; that is, preparatory evidence to support your claim that something like the approach you are proposing in your plan and implementing in your report 2 would positively influence business metrics that are relevant to the client (say, Spotify). Completing the report will require working with data, analysing the data using some method or modelling framework (as I have said above, a regression technique for predictive analytics.) then providing a statement of possible returns should the pilot be expanded to a full project per your suggested investment. I am defining the term investment broadly – the financial and organisational resources needed to expand your pilot to a full project. A Precursor to the Project Report Furthermore, think of the plan as a precursor to your report – in fact, the project plan should form the first part of the report you later submit. The report is comprehensive – sketching out all aspects of the project from beginning to end. Key sections include a statement of the business problem you are trying to solve, the proposed method of analysis, the results and interpretations, etc. Some of these sections will be first developed in the proposal and later repeated in the full report. The full report may have the following sections. 1. Background 2. Business problem 3. Proposed solution/method of analysis 4. Data 5. Analysis plan 6. Results and interpretation 7. Strategy/recommendations (and business case) 8. Conclusion Key Sections in the Project Plan The plan might include the first three sections planned for the report plus an initial sketch of the analysis plan. To be specific, the plan might include the following. 1. Background 2. Business problem 3. Proposed solution/method of analysis 4. Analysis plan 5. Project directions The background section introduces the plan, placing emphasis on the broad context and highlighting the specific domain of interest (accounting/financial analytics, social media analytics, etc.). To add realism to your project plan and report, you might adopt the perspective of a business analyst and/or consultant pitching your solution at a target business (i.e., a real or hypothetical business – probably better to have a real business in mind). The examples offered earlier might have an established recruitment company (e.g., Blenheim Partners) and a fast-service restaurant (Taco Bell) as the target businesses/clients, respectively. The business/client might be described and the potential benefits of doing the project could emphasised. The general expectation is some key business process and/or management decision will be improved through the use of analytics – keep in mind the project plan is also a marketing document (Sant 2019). Perhaps the most important part of the plan is specifying the business problem. This may also be the most challenging part of completing the project plan and report. What is the problem you are trying to solve? A starting point might be to specify an outcome the target business/client is trying to achieve. Addressing the following questions may help. (1) What is the specific outcome the 3 business is trying to influence or understand, or both? (2) How is it manifest or measured? (3) What level of that outcome variable has the business achieved to date, and what is the desired level of the outcome variable? (4) What are the likely factors or other variables that are potentially related to the outcome of interest, and how are they measured? The section on the proposed solution/method of analysis – really, the analytical “line of attack” – should focus on the method of analysis – a particular method of business analytics suitable for the problem you have outlined/specified. Describe the method in some detail. Highlight key steps and/or issues in its implementation. For example, you might write-out the regression model in formal fashion with reference to the specific output and feature variables you will select for study. Why these variables? What is the nature of these variables? Furthermore, this section should also outline the data requirements needed to solve the business problem. In summary, this section should be specific about the proposed method of analysis and the data to be used. The analysis plan is intended as a step-by-step guide to implementing the proposed method with some emphasis of the basic assumptions of the method. For example, a regression analysis might involve several key steps, including the appropriateness of the data for the planned application, the specification of the outcome and input variables, and some assessment of the how the basic assumptions of regression might be satisfied. This section may be difficult to write if only because you are asked to anticipate the application of a particular method to a given dataset – without necessarily having much experience with the method and/or knowledge of the data. As a result, this section may have somewhat of a textbook feel. The final section, project directions, should sketch the “next steps” for your plan of action for completing the project and report. What needs to be done to complete the project, when will it be done, what will the outcomes and/or deliverables be – what form will they take. Often, the reporting of projects of the type you will complete in this course takes the form of a slide deck (PowerPoint) and a face-to-face presentation. This is my expectation for the form of the full report – perhaps specify that the outcomes will take this form in this section of the proposal. Also, in this section remember to link back to the strategic value of the project – remind the client why the project is important and should be done now. Submission Guidelines The project plan is worth 20 percent of your score in the course. I will use something like the indicative marking scheme shown below to score the proposals. I will rate each section on a 1-7 scale and provide an overall score out of 20. Take a rating of one to indicate “poor” (failing to demonstrate basic requirements) and a rating of seven to indicate “excellent” (showing substantial insight and originality). My expectation is you write in the range of 2,000 words for the project plan. Include a reference list if needed. Getting Started You might consider the following steps in getting started on the project plan. Probably the most important step is setting scope – do so with a vision for your overall project in mind. 1. Maybe start by choosing a business of interest to you (this Semester, you are encouraged to think about music streaming services such as Spotify or Apple Music) 2. Think about some of the core business problems that business would address that might be solved by the use of analytics 4 3. Think about what method might be particularly suited to the problem(s) you are considering for study 4. Do some background reading on the business and on the method (taking notes as you do) 5. Write the problem statement 6. Write up your notes on the method in a form suitable for the proposed solution section 7. The notes may also help you sketch out the key points of the analysis plan Finally, think of these notes as a guide only. They communicate my broad expectations – but you may think of different and better ways to approach writing the proposal.
欢迎咨询51作业君
