| Field | Details |
|---|---|
| Subject Code | DATA6000 |
| Subject Name | Capstone: Industry Case Studies |
| Assessment Title | Project Report |
| Assessment Type | Individual Report and Pitch |
| Assessment Length | 2000 Words (+/-10%) |
| Weighting | 35% Report / 15% Pitch |
| Total Marks | 50 |
| Submission | Turnitin and in class |
| Due Date | In class and Friday Week 12 |
To synthesise your learnings from the Business Analytics course into a report, you need to undertake an analytics project and prepare an industry research report.
Objective: Your objective is to develop a solution that must:
This assessment aims to achieve the following subject learning outcomes:
In your report, please follow the below structure. The words per section are only a suggestion.
1. Executive Summary (100 words)
2. Industry Problem (300 words)
3. Data Processing and Management (400 words)
4. Data Analytics Methodology (400 words)
5. Visualisation and Evaluation of Results (300 words)
6. Recommendations (400 words)
7. Data Ethics and Security (100 words)
8. Elevator Pitch (3 Minutes)
Academic Integrity and Conduct Policy https://www.kbs.edu.au/admissions/forms-and-policies
KBS values academic integrity. All students must understand the meaning and consequences of cheating, plagiarism and other academic offences under the Academic Integrity and Conduct Policy.
Please read the policy to learn the answers to these questions:
Late Submission of Assignments
| Number of Days Late | Penalty |
|---|---|
| 1* – 9 days | 5% per day for each calendar day late deducted from the total marks available. |
| 10 – 14 days | 50% deducted from the total marks available. |
| After 14 days | Assignments submitted more than 14 calendar days after the due date will not be accepted and the student will receive a mark of zero for the assignment(s) unless special consideration, reasonable adjustment or an alternative factor related to compassionate circumstances is approved and applied. |
Assignments submitted at any stage within the first 24 hours after the deadline will be considered to be one day late and therefore subject to the associated penalty.
Length Limits for Assessments Penalties may be applied for assessment submissions that exceed prescribed limits.
Study Assistance Students may seek study assistance from their local Academic Learning Advisor or refer to the resources on the MyKBS Academic Success Centre page. Further details can be accessed at https://elearning.kbs.edu.au/course/view.php?id=1481
| Traffic Light | Amount of Generative AI Usage | Evidence Required | This Assessment |
|---|---|---|---|
| Level 1 – Prohibited | No GenerativeAI allowed. This assessment showcases your individual knowledge, skills and/or personal experiences in the absence of Generative AI support. | The use of generative AI is prohibited for this assessment and may potentially result in penalties for academic misconduct, including but not limited to a mark of zero for the assessment. | |
| Level 2 – Optional | You may use GenerativeAI for research and content generation that is appropriately referenced. This assessment allows you to engage with Generative AI as a means of expanding your understanding, creativity, and idea generation in the research phase of your assessment and to produce content that enhances your assessment. | The use of GenAI is optional. Your collaboration with GenerativeAI must be clearly referenced. You must include an appendix that documents your GenerativeAI collaboration including all prompts and responses. | ✓ |
| Level 3 – Compulsory | You must use GenerativeAI to complete your assessment. This assessment fully integrates Generative AI, allowing you to harness the technology's full potential in collaboration with your own expertise. | You will be taught how to use generative AI and assessed on its use. Your collaboration must be clearly referenced and an appendix documenting all prompts and responses must be included. |
| Standards for this Task | Points | Feedback |
|---|---|---|
| Problem Statement — Clear executive summary; clear description of the industry problem; clear description of data processing and management; well-researched project with accurate and relevant referencing. | /10 | |
| Results, Analysis & Recommendations — Extensive coverage of analytics methodology including an appendix. Multiple data sources used effectively. Clear forward-looking outcomes. Extensive discussion of project recommendations. Clear outline of privacy, legal, security and ethical considerations. Analytics model file uploaded to file dropbox (if missing, marks for this section = zero). For a higher grade: original and challenging business problem; multiple, technically sophisticated analytics methods. | /20 | |
| Report — Appropriate structure; ten or more relevant references; in-text references related to paragraphs; use of GenerativeAI in accordance with Traffic Lights; report uploaded to Turnitin and analytics model to file Dropbox. | /5 | |
| Elevator Pitch — Appropriate arguments to convince audience; understanding of competitive benefit; ability to answer questions; project pitched must match Assessment 2 and Assessment 3 report. | /15 | |
| Total | /50 |
Note: This report is provided as a sample for reference purposes only. For further guidance, detailed solutions, or personalized assignment support, please contact us directly.
Predicting Customer Churn in the Telecommunications Industry Using Predictive Analytics
1. Executive Summary (Sample – 100 Words)
Customer churn is a major challenge in the telecommunications industry because losing customers directly impacts revenue and profitability. This project uses predictive analytics to identify customers who are likely to leave a telecom company. A dataset containing customer demographics, service subscriptions, billing information, and contract details was analysed using descriptive and predictive analytics techniques. Logistic Regression and Decision Tree models were applied to predict churn behaviour. The analysis revealed that contract type, monthly charges, internet service, and customer tenure significantly influence churn. Recommendations include introducing personalised retention strategies, loyalty programs, and proactive customer support to reduce churn and improve long-term customer retention.
2. Industry Problem (Sample – 300 Words)
The telecommunications industry is highly competitive, with companies constantly trying to attract and retain customers. One of the biggest challenges faced by telecom providers is customer churn, which occurs when customers discontinue their services and move to competitors. High churn rates reduce profitability because acquiring new customers is significantly more expensive than retaining existing ones.
In recent years, telecom companies have accumulated large amounts of customer data through billing systems, online interactions, and service usage records. This creates opportunities for businesses to apply data analytics techniques to understand customer behaviour and improve retention strategies.
The business problem addressed in this project is:
“How can predictive analytics help identify customers who are likely to churn in a telecommunications company?”
Solving this problem is important because customer retention directly affects revenue growth, customer lifetime value, and market competitiveness. By predicting churn early, telecom companies can take proactive actions such as personalised promotions, loyalty rewards, or targeted customer support.
Data analytics is highly suitable for this problem because customer churn patterns can be identified using historical customer behaviour. Descriptive analytics helps uncover trends and patterns in customer demographics and service usage, while predictive analytics can forecast the likelihood of churn.
The availability of publicly accessible telecom churn datasets strongly influenced the selection of this project topic. The dataset provided structured customer information suitable for applying machine learning models. Furthermore, customer churn prediction is widely recognised as a real-world business problem, making it highly relevant for industry application.
This project demonstrates how data-driven decision-making can help businesses improve customer satisfaction and reduce operational losses through predictive analytics.
3. Data Processing and Management (Sample – 400 Words)
The dataset used for this project was the Telco Customer Churn dataset obtained from Kaggle. The dataset contains customer information including demographics, contract details, internet services, billing methods, monthly charges, tenure, and churn status.
The dataset includes approximately 7,000 customer records with 21 variables. The target variable was “Churn,” which indicates whether a customer left the company.
Data Relevance
The dataset is highly relevant because it contains variables directly associated with customer retention behaviour. Features such as contract type, payment method, tenure, and monthly charges are valuable indicators for churn prediction.
Descriptive Analytics
Descriptive analytics techniques were used to identify customer behaviour trends. Data visualisations such as bar charts, pie charts, histograms, and correlation heatmaps were used to analyse:
The descriptive analysis showed that customers with month-to-month contracts and higher monthly charges had higher churn rates.
Predictive Analytics
Predictive analytics techniques were applied using machine learning models. Logistic Regression and Decision Tree Classification were selected because they are commonly used classification algorithms suitable for churn prediction problems.
The dataset was divided into training and testing datasets using an 80:20 split ratio.
Data Cleaning and Preparation
Several preprocessing steps were conducted:
Python libraries including Pandas, NumPy, Matplotlib, Seaborn, and Scikit-learn were used for data preparation and analysis.
Data Mining Process
The data mining process involved:
The structured preprocessing improved model accuracy and reliability.
4. Data Analytics Methodology (Sample – 400 Words)
This project followed the CRISP-DM (Cross Industry Standard Process for Data Mining) methodology. CRISP-DM is widely used in business analytics because it provides a structured framework for solving business problems using data.
1. Business Understanding
The first phase involved identifying the business issue of customer churn in the telecommunications industry. The objective was to predict customers at risk of leaving.
2. Data Understanding
The dataset was explored using descriptive statistics and visualisations. Variables such as monthly charges, contract type, and tenure were identified as important churn indicators.
3. Data Preparation
Data preparation included cleaning missing values, transforming categorical variables, and standardising features for machine learning algorithms.
4. Modelling
Two predictive models were selected:
Logistic Regression
This model was chosen because it is interpretable and effective for binary classification problems.
Decision Tree Classifier
This model was selected because it provides visual decision rules and handles non-linear relationships effectively.
The models were trained using historical customer data.
5. Evaluation
The models were evaluated using:
The Decision Tree model achieved higher predictive performance compared to Logistic Regression.
6. Deployment
Although full deployment was outside the project scope, the predictive model could be integrated into customer relationship management (CRM) systems to identify at-risk customers in real time.
Rationale for Methodology
CRISP-DM was chosen because it aligns business understanding with technical analytics implementation. The methodology ensures systematic project development and supports practical industry applications.
Appendix (Sample Content)
5. Visualisation and Evaluation of Results (Sample – 300 Words)
Several visualisations were developed to communicate insights effectively.
Descriptive Insights
Churn Distribution Chart
The visualisation showed that approximately 27% of customers churned.
Contract Type Analysis
Customers with month-to-month contracts had significantly higher churn rates compared to customers with yearly contracts.
Monthly Charges Histogram
Higher monthly charges were associated with increased customer churn.
Predictive Analytics Results
The Decision Tree model achieved an accuracy score of 82%, while Logistic Regression achieved 79%.
The confusion matrix demonstrated that the Decision Tree model more accurately classified churn and non-churn customers.
Evaluation
The visualisations successfully highlighted important churn factors and improved the interpretability of the predictive models.
Python libraries including Matplotlib and Seaborn were effective for generating clear business visualisations. Scikit-learn provided reliable machine learning capabilities.
However, some limitations existed:
Despite these limitations, the analytics models provided meaningful insights that could support customer retention strategies.
6. Recommendations (Sample – 400 Words)
Based on the analysis, several recommendations are proposed.
1. Introduce Loyalty Programs
Customers with month-to-month contracts showed higher churn probability. Telecom companies should provide discounts and loyalty rewards for long-term contracts.
2. Personalised Retention Campaigns
Predictive analytics can identify high-risk customers before they leave. Companies should use personalised offers and targeted communication strategies.
3. Improve Customer Support
Poor customer service often contributes to churn. Businesses should strengthen customer support channels and resolve complaints quickly.
4. Reduce High Monthly Charges
Customers paying higher monthly fees showed increased churn rates. Flexible pricing plans and bundled services may improve retention.
5. Implement Real-Time Analytics Systems
Companies should integrate predictive models into CRM systems to monitor customer behaviour continuously.
Communication to Diverse Stakeholders
For executives:
For marketing teams:
For technical teams:
Limitations
The project faced several limitations:
Role of Data Analytics
Data analytics plays a critical role in modern business decision-making. Predictive models enable proactive business strategies instead of reactive responses.
Future Improvements
Future projects could include:
These technologies would improve prediction accuracy and business intelligence capabilities.
7. Data Ethics and Security (Sample – 100 Words)
Data ethics and security are essential in customer analytics projects. Customer information must be stored securely and comply with privacy regulations such as GDPR and Australian Privacy Principles. Sensitive customer data should be anonymised and protected from unauthorised access.
Visualisations and predictive outputs must remain accurate and transparent to avoid misleading stakeholders. Ethical concerns also include algorithmic bias and fairness in predictive modelling.
Future analytics systems should prioritise responsible AI practices, transparency, informed consent, and cybersecurity protections to ensure ethical and trustworthy business analytics implementation.
8. Elevator Pitch Script (3 Minutes Sample)
“Good afternoon everyone.
Customer churn is one of the largest financial challenges in the telecommunications industry. Acquiring a new customer costs significantly more than retaining an existing one.
Our project uses predictive analytics to identify customers who are likely to leave a telecom company before they actually churn.
Using a real-world telecom dataset, we applied descriptive analytics to uncover patterns in customer behaviour and predictive machine learning models including Logistic Regression and Decision Trees to forecast churn risk.
Our analysis revealed that customers with month-to-month contracts, higher monthly charges, and shorter tenure are more likely to leave.
The Decision Tree model achieved over 82% prediction accuracy, demonstrating strong business value.
Based on these findings, we recommend loyalty programs, personalised retention campaigns, improved customer service, and real-time analytics integration into CRM systems.
This solution enables telecom companies to proactively retain customers, reduce revenue loss, and improve long-term profitability.
In the future, this model can be enhanced using artificial intelligence, customer sentiment analysis, and real-time analytics technologies.
Thank you.”
Suggested Visuals for Report
You can include:
Suggested Tools
Sample Academic References
Get original papers written according to your instructions and save time for what matters most.
