Description

Purpose

There are two parts in this assignment.

•    Part A provides you with opportunities to learn a range of machine learning methods and Python skills (GLO1 & ULO1) and apply your digital literacy to research and develop a machine learning solution (GLO3, GLO5, and ULO2). By completing this task, you will gain knowledge and skills in selecting and applying one or more appropriate machine learning algorithm(s) to develop and evaluate a machine learning solution and interpret the outcomes.

•    In Part B, you will report your application of machine learning and make recommendations to the business and management audience. By completing this task, you will gain ability to explain and justify machine learning options and discuss their pros and cons to the business audience.

Context/Scenario

VSNeoBank, a fictitious business, is a digital-only banking platform headquartered in Melbourne, Victoria. They operate solely online to provide banking services to customers in all states of Australia. For a quick introduction to neobanks, please refer to the following articles:

https://www.forbes.com/advisor/banking/what-is-a-neobank/

https://www.infochoice.com.au/news/what-is-a-neobank/

Recently, the  company  has  noted  an  increased  customer  attrition  rate. You  have  been given  a  dataset, generated from their  customer  and transaction  databases.  Ms.  Emma  Hoang,  Head  of VSNeoBank  Data Analytics, is keen to explore their business data to better understand their customers and improve customer retention.

Datasets provided:

•    VSNeoBank.csv

•    VSNeoBank_competition.csv

•    Data description

You are required to explore the first dataset VSNeoBank.csv and develop and test machine learning options(s) using Python. You are also required to develop two reports. The first report (Part A) should present your machine learning findings to Ms. Emma Hoang, Head of Data Analytics. This report should detail your approach to exploring the dataset, the machine learning techniques used, and your findings. Your findings should be supported by relevant visualizations and statistical analysis. Based on your machine learning findings, you are then required to develop a consultancy report (Part B) for Mr. Martin Perry, Head of Customer Experience at VSNeoBank. This report should provide actionable recommendations that can help the bank improve its customer experience and retention. The report     should include a summary of your findings, key insights, and recommendations. You should also discuss the limitations of your approach and any potential areas for future improvements. See further details in the section Specific Requirements below.

Optional: You are invited to deploy the model and apply it to a second dataset (VSNeoBank_competition.csv) to demonstrate your ability to save and deploy the model on unseen data. The model with the highest accuracy score will receive a small prize.

The source of the dataset will be provided after the assignment has been returned.

Specific Requirements

You are required to:

•    Develop your business and data understandings.

•    Prepare and explore the provided dataset, cleanse and pre-process data as needed.

•    Undertake machine learning model development, evaluation and selection. Two models should be developed, tested and compared.

•    Develop two reports:

•   The first report (Part A) should present your machine learning findings to Ms. Emma Hoang, Head of Data Analytics, VSNeoBank.

•   The second consultancy  report  (Part  B) to  Mr.  Martin  Perry,  Head of Customer  Experience, VSNeoBank.

•    Format and present your report professionally. Two sample report templates are provided under Assessment Resources.

•   Correctly use the APA7 style of referencing, and include in-text citations when quoting, referring to, summarising, or paraphrasing from any source:

https://www.deakin.edu.au/students/studying/study-support/referencing

Deliverables:

Part A. Case Study Report

Part A.1 Machine Learning Solution

•    A cover page (not included in the word count) that includes:

•    Report Title

•    Unit code and name

•    Student name and student ID

•    A table of contents (not included in the word count)

•    An executive summary of max. 200 words is required (included in the word count).

•   The report should include:

1.    Introduction:

•   The business problem to address and its business context using the Business Analysis Core Concept Model (BACCM) framework1  .

2.   Approach:

•   Overview of the machine learning approach for the project and selection criteria

3.    Data preparation :

•   Data sources and contents, data cleansing and pre-processing, and any challenges

4.    Exploratory data analysis (EDA):

•   Statistical analysis and visualisation

•   Key insights gained from EDA

•   Factors that influence the model(s)

5.    Model development and evaluation:

•   Machine learning models developed, tested and performance metrics

•   Model comparison based on the selection criteria

6.    Results and interpretation:

•   Presentation of the proposed solution

•   Interpretation and discussion of test/validation results obtained from the test/validation data

•   Presentation of the results obtained from the deployment data

7.   Technical recommendations :

•   Summary  of  the  development  and  testing  environment,  such  as  software  libraries, programming languages, and computing resources used.

•   Suggestions for model deployment

•   Suggestions for maintenance of accuracy and relevance over time

8.    Recommendations  for  further  improvements  and  potential  business  applications  and implications

•    References (not included in the word count)

•    Optional appendices (not included in the word count – not subject to assessment), such as additional technical details, supplementary figures and tables

Part A.2 Files

•    A python notebook with detailed comments to guide the deployment team

•    A csv deployment file with your predicted labels

Part B. Business report

•    A cover page (not included in the word count) that includes:

•    Report Title

•    Unit code and name

•    Student name and student ID

•    A table of contents (not included in the word count)

•    An executive summary of max. 100 words is required (included in the word count).

•    Use APA7 style of referencing and include in-text citations:

https://www.deakin.edu.au/students/studying/study-support/referencing

Learning Outcomes

This task allows you to demonstrate your achievement towards the Unit Learning Outcomes (ULOs) which have been aligned to the Deakin Graduate Learning Outcomes (GLOs). Deakin GLOs describe the knowledge and capabilities graduates acquire and can demonstrate on completion of their course. This assessment task is an important tool in determining your achievement of the ULOs. If you do not demonstrate achievement of the ULOs you will not be successful in this unit. You are advised to familiarise yourself with these ULOs and

GLOs as they will inform you on what you are expected to demonstrate for successful completion of this unit.

The learning outcomes that are aligned to this assessment task are:

Submission

You must submit your assignment in the Assignment Dropbox in the unit CloudDeakin site on or before the due date. The submission must include two files:

•   Two (2) report documents. Name your documents using the following syntax: <your surname_your first name_your Deakin student ID number_[unitcodeA1].doc (or ‘ .docx’). For example,                   ‘MIS710A1_Jones_Barry_123456789_MIS710A2 ReportA.doc’ and                                                           ‘MIS710A1_Jones_Barry_123456789_MIS710A2 ReportB.doc’

•    One (1) Python notebook

•    One (1) csv file with labels VSNeoBank_deploy_predlabels.csv

Submitting a hard copy of this assignment is not required. You must keep a backup copy of every assignment you submit until the marked assignment has been returned to you. In the unlikely event that one of your assignments is misplaced you will need to submit your backup copy.

Any work you submit may be checked by electronic or other means for the purposes of detecting collusion and/or plagiarism and for authenticating work.

When you submit an assignment through your CloudDeakin unit site, you will receive an email to your Deakin email address confirming that it has been submitted. You should check that you can see your assignment in the Submissions view of the Assignment Dropbox folder after upload and check for, and keep, the email receipt for the submission.

Marking and feedback

The marking rubric indicates the assessment criteria for this task. It is available in the CloudDeakin unit site in the Assessment folder, under Assessment Resources. Criteria act as a boundary around the task and help specify what assessors are looking for in your submission. The criteria are drawn from the ULOs and align with the GLOs. You should familiarise yourself with the assessment criteria before completing and submitting this task.

Students who submit their work by the due date will receive their marks and feedback on CloudDeakin 15 working days after the submission date.

Extensions

Extensions can only be granted for exceptional and/or unavoidable circumstances outside of your control.

Requests for extensions must be made by 12 noon on the submission date using the online Extension Request form under the Assessment tab on the unit CloudDeakin site. All requests for extensions should be supported by appropriate evidence (e.g., a medical certificate in the case of ill health).

Applications for extensions after 12 noon on the submission date require University level special consideration and these applications must be must be submitted via StudentConnect in your DeakinSync site.

Late submission penalties

If you submit an assessment task after the due date without an approved extension or special consideration, 5% will be deducted from the available marks for each day after the due date up to seven days*. Work submitted more than seven days after the due date will not be marked and will receive 0% for the task. The Unit Chair may refuse to accept a late submission where it is unreasonable or impracticable to assess the task after  the  due  date.  *'Day'  means  calendar  day  for  electronic  submissions  and  working  day  for  paper submissions.

An example of how the calculation of the late penalty based on an assignment being due on a Thursday at 8:00pm is as follows:

•    1 day late: submitted after Thursday 11:59pm and before Friday 11:59pm– 5% penalty.

•    2 days late: submitted after Friday 11:59pm and before Saturday 11:59pm – 10% penalty.

•    3 days late: submitted after Saturday 11:59pm and before Sunday 11:59pm – 15% penalty.