ABC Trading Ltd. (ABC) is a garment
manufacturing company headquartered in Hong Kong. Its manufacturing facilities
are spread all over Asia, in countries such as China, Malaysia, Thailand, the
Philippines, and Vietnam. It produces over 30 million garments, mostly ladies’
apparel, annually.
The garment industry is one of the most
labour-intensive industries. A number of human factors, such as workmanship and
human errors, can easily affect product quality. To assure the quality of
products, ABC provides training for operators to enhance their job skills in
different operations such as fabric spreading, cutting, sewing, and finishing.
It also tracks operators’ performance, which is used to determine salary
increment, bonuses and promotions.
Currently, ABC does not have a systematic
approach in reviewing the effectiveness of the training programme. In addition,
it is found that the number of product quality issues is increasing, causing a
large amount of product rework cost. ABC would like to use data mining to
examine factors associated with good operator performance so that its training
resources can be more effectively allocated for quality improvement. The
dataset – ABC.csv, relates to the training programme offered by ABC and the
description of the attributes is listed in Table 1.
Table 1. Description of ABC.csv
Attribute Name |
Description |
Labels / Values |
OpID |
Operator identity |
Unique number |
Gender |
Gender |
0 = “male” 1 = “female” |
Edu |
Education level |
0 = “secondary & below” 1 = “above secondary” |
Tedu |
Type of education |
0 = “academic” 1 = “vocational” |
Wexp |
Prior work experience |
0 = “2 years or less” 1 = “more than 2 years” |
Twexp |
Type of prior work experience |
0 = “office work” 1 = “factory work” |
Trg |
Completion of operator training |
0 = “no” 1 = “yes” |
Age |
Age |
Integer measurement |
Trgper |
Training performance index |
Integer measurement |
Qual |
Qualitative skills index |
Integer measurement |
Quant |
Quantitative skills index |
Integer measurement |
Wkrate |
Weekly rate |
Integer measurement |
Mcdays |
Average days of sick leave taken per year |
Integer measurement |
Yrsco |
Years with company |
Integer measurement |
Opper |
Operator performance |
0 =
“not-so-good” 1 = “good” |
Question:
(a) With reference to the pre-modelling,
modelling, and post-modelling stages of data mining, explain the key tasks to
be performed when ABC applies data mining to achieve its business objective.
Your answer must be less than 300 words.
(30 marks)
(b) For this task, you are required to
analyse variables that are worth exploring in order to detect interesting
patterns in the data. With the aid of diagrams, perform the following data
exploration tasks using appropriate data visualisation techniques (one technique
for each task):
(i)
Univariate data exploration (10 marks)
(ii)
Bivariate data exploration (10 marks)
(iii) Multivariate data exploration (10 marks)
(c) Discuss the findings of each data
exploration task you performed in part (b). (24 marks)
(d) Identify one (1) variable that contains
missing values. As part of your data preparation consideration, what is your
suggestion for handling the missing values of this variable? Justify your suggestion. (6 marks)
Answers should not exceed 1800 words or 10
pages (including all the relevant graphs and tables)
Please use IBM SPSS Modeler
Get Free Quote!
363 Experts Online