TTTU6214

FINAL ASSESSMENT TU6214

INSTRUCTION: Two (2) Tasks (Task 1 and Task 2) are available for your final assessment. Read
these two tasks and choose the one which is the best for you or the one you are most
interested in. You can write your answer in English or Bahasa Melayu. Choose only ONE task
to complete.

TASK 1: CHURN ANALYSIS (50 marks)

According to padding.com, churn analysis is the evaluation of a company's customer loss rate
to reduce it. Also referred to as customer attrition rate, churn can be minimized by assessing
your product and how people use it. In other words, churn is the rate customers leave your
business and take their subscription money elsewhere.

This assignment gives you a company's customer data (L1customerdatav02.xlsx). Your job is
to use a machine learning model that predicts which customer will churn. Students are
advised to read an article on Customer Churn Analysis to better understand the problem and
the given data. There are many resourceful articles that Googling can find. One of the best is
published by Paddle.com (https://www.paddle.com/resources/customer-churn-analysis)

1. The data engineer believed that there were inconsistencies in the data. You are required
to identify and resolve data quality issues.

a) Examine the sample data in the '(L1customerdatav02.xlsx)' Excel workbook. Identify data
quality issues and provide recommendations to resolve them. Note that the given data is not
the data for the entire business but is a representative sample subset. Fix the issues and write
an informal memo to your supervisor to document your findings (the errors) and how you
have fixed them.
(5 marks)

b) One of the issues raised by the management is their concern about breaching any privacy
law in implementing churn analysis, as the analysis output will be used in their loyalty
marketing program. Your supervisor believes that the company is not breaking any rules for
analytics purposes if any personally identifiable data has been masked or removed. Your task
is to review whether churn analysis with anonymization of the customer (data masking)
breaches the Personal Data Protection Act 2010 (PDPA). Write an informal memo to your
supervisor (English or Malay) lead to document your findings and recommendations.

Copy of the PDPA can be downloaded from:

(10 marks)

1
 TTTU6214

2. Use the clean data to develop the churn prediction model.

a. Develop a predictive model comprising factors to predict mortality using K-NN. Report
the accuracy, AUC, and F-Measure of the model.
(5 marks)

b. Develop a classifier model using the Decision Tree operator. Report the accuracy, AUC,
and F-Measure of the model.
(5 marks)

c. Improve the results by using any machine learning method. Explain how you have
improved the quality of the results. Please list the newly added operators/functions and
the latest results (i.e., Accuracy, AUC, and F-Measure).
(10 marks)

d. Improve the results of the generated K-NN model. Explain how you have improved the
quality of the results. Please list the newly added operators/functions and the latest
results (i.e., Accuracy, AUC, and F-Measure).
(10 marks)

e. Explain whether your results are overfitting or underfitting? Justify your answer.
( 5 marks)

2
 TTTU6214

TASK 2 (50 marks)

1. The dataGDP.xlsx file contains Malaysia's gross domestic product per capita (GDP) from
1960 to 2020.
a. Visualize Malaysia's GDP from 1960 to the most recent years using the most
appropriate diagram.
(4 marks)

b. As an analyst, write a report for the new Prime Minister describing the
information in the diagram you just developed for 1a. Your report must
include a visualization (i.e. diagram/graph) to support your interpretation of
the data. Comment on the appearance and strength of any trend.

(10 marks)

2. Based on the diagram developed by you, explain whether the relationship between the
year and GDP appears to be deterministic or involves randomness. Justify your answer.
(6 marks)

3. Use a spreadsheet to forecast Malaysia's GDP for 2020. Use TWO forecasting techniques
to forecast the 2020 GDP. Report the results. Please upload the spreadsheet for
evaluation.

(10 marks)

4. Your supervisor gave you another file (militarySpending.xlsx). He asked you to find
insights from the data on Malaysia's military spending trend. Explain the relationship
between Malaysia's military spending and GDP and compare the patterns to other
ASEAN countries' military expenditures, such as Singapore, Indonesia and Thailand.

Write an informal memo to your supervisor (English or Malay) about your findings. Use
the correct chart or graph that can show/visualize connections between data and
patterns of behaviour.

(20 marks)

SELAMAT MAJU JAYA