A VISUALIZED ANALYSIS ON STUDENT PERFORMANCE BY USING CLOUD

BUSINESS INTELLIGENCE
BY
CHEAH CHENG QIAN

A REPORT
SUBMITTED TO
Universiti Tunku Abdul Rahman
in partial fulfillment of the requirements
for the degree of
BACHELOR OF INFORMATION SYSTEMS (HONOURS) BUSINESS INFORMATION
SYSTEMS
Faculty of Information and Communication Technology
(Kampar Campus)

JUNE 2022
A VISUALIZED ANALYSIS ON STUDENT PERFORMANCE BY USING CLOUD
BUSINESS INTELLIGENCE
BY
CHEAH CHENG QIAN

A REPORT
SUBMITTED TO
Universiti Tunku Abdul Rahman
in partial fulfillment of the requirements
for the degree of
BACHELOR OF INFORMATION SYSTEMS (HONOURS) BUSINESS INFORMATION
SYSTEMS
Faculty of Information and Communication Technology
(Kampar Campus)

JUNE 2022

i
 UNIVERSITI TUNKU ABDUL RAHMAN

REPORT STATUS DECLARATION FORM

Title: A VISUALIZED ANALYSIS ON STUDENT PERFORMANCE BY

__________________________________________________________
USING
USING CLOUD BUSINESS INTELLIGENCE

Academic Session: 2022

I CHEAH CHENG QIAN

(CAPITAL LETTER)

declare that I allow this Final Year Project Report to be kept in

Universiti Tunku Abdul Rahman Library subject to the regulations as follows:
1. The dissertation is a property of the Library.
2. The Library is allowed to make copies of this dissertation for academic purposes.

Verified by,

_________________________ _________________________
(Author’s signature) (Supervisor’s signature)

Address:
No, 32 Laluan Pinji Wani,
__________________________
Taman Pinji Wani
__________________________ Ts Dr Chan Lee Kwun
31650 Ipoh Perak
__________________________ Supervisor’s name

Date: ____02 September 2022____ Date: __02 September 2022___

ii
 Universiti Tunku Abdul Rahman
Form Title : Sample of Submission Sheet for FYP/Dissertation/Thesis
Form Number: FM-IAD-004 Rev No.: 0 Effective Date: 21 JUNE 2011 Page No.: 1 of 1

FACULTY/INSTITUTE* OF INFORMATION AND COMMUNICATION TECHNOLOGY

UNIVERSITI TUNKU ABDUL RAHMAN

Date: 02 September 2022

SUBMISSION OF FINAL YEAR PROJECT /DISSERTATION/THESIS

It is hereby certified that Cheah Cheng Qian (ID No: 17ACB04662 ) has completed this final year
project/ dissertation/ thesis* entitled “A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence” under the supervision of Ts Dr Chan Lee Kwun (Supervisor) from the
Department of Information Systems, Faculty/Institute* of Information and Communication
Technology , and ____________ (Co-Supervisor)* from the Department of ___________________,
Faculty/Institute* of __________________________.

I understand that University will upload softcopy of my final year project / dissertation/ thesis* in pdf
format into UTAR Institutional Repository, which may be made accessible to UTAR community and
public.

Yours truly,

____________________
(CHEAH CHENG QIAN)

*Delete whichever not applicable

iii
 DECLARATION OF ORIGINALITY

I declare that this report entitled “A VISUALIZED ANALYSIS ON STUDENT

PERFORMANCE BY USING CLOUD BUSINESS INTELLIGENCE” is my own work
except as cited in the references. The report has not been accepted for any degree and is not
being submitted concurrently in candidature for any degree or other award.

Signature : _________________________

Name : Cheah Cheng Qian

Date : 02 September 2022

iv
 ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to several individuals and organizations for
supporting me throughout my Graduate study. First, I wish to express my sincere gratitude to
my supervisor, Ts Dr Chan Lee Kwun, for her enthusiasm, patience, insightful comments,
helpful information, practical advice, and unceasing ideas that have always helped me
tremendously in my research and writing of this thesis. Her immense knowledge, profound
experience, and professional expertise in the field of Business Intelligence has enabled me to
complete this research successfully. Without her support and guidance, this project would not
have been possible. I could not have imagined having a better supervisor in my study.

To a very special person in my life, Yeoh Pei Ying, for her patience, unconditional support,
and love, and for standing by my side during hard times. Finally, I must say thanks to my
parents and my family for their love, support, and continuous encouragement throughout the
course.

Finally, a big thanks to Universiti Tunku Abdul Rahman especially to the Faculty of
Information Technology for giving me a chance to study here.

v
 ABSTRACT

During in this state of global pandemic, it is claimed that there is a rise in trend of poor
academic performance among student in certain courses or other specific skills. In this matter,
due to this uprising yet concerning issue, it is crucial that the deteriorations of students’
academic performance seek a proper solution. Previous research has primarily relied on the
traditional and simple method to analyze students’ performance. However, the results and
outcomes are not too accurate due to lack of limited data as well as analysis of students’ CLO
performance. Therefore, possible factors have examined to have certain degree of relevance in
affecting students' performance in certain skills or courses. The aim of this project is to gain
deeper understanding and insights to find out which skills or courses where most students can’t
perform well and to propose a solution into solving the issue by implementing analysis tools
using software tools like Python and MS Power BI. Hence, a comparison of business
intelligence tools will be carried out to choose an appropriate BI tool to conduct the analysis
due to vast amount of software which available in the IT market. Next, a simple prototype of a
system is developed to conduct the analysis alongside a business intelligence tool to graphically
display and visualize the analysis results along with graphs on the existing dashboard. To
analyze students’ performance, a sample data of students’ results is created to check its CLO
through specific and measurable statements in which defines students’ knowledge, skills, and
attitudes by demonstrating by the completion of a course. Cognitive, Practical and Affective
acts as an indicator to see which student did not perform well in which components in a course
and indicates what a student should understand and be able to do after learning. The dashboards
developed will be used as a module to analyze students’ performance. After analyzing the
findings through graphical objects, formulations and recommendations, the results and
outcomes of the data analysis are shown and presented and then described based on the finding.
Recommendations to overcome the issue is proposed and conclusions are drawn for future
research opportunities as well as data collection. The results of this project will help us gain a
better insight and understanding on the academic performance of students as well as knowing
which area of weaknesses are they into so that students in higher education could be able to
have a huge leap in academic performance.

vi
 TABLE OF CONTENTS

TITLE PAGE i
REPORT STATUS DECLARATION FORM ii
FYP THESIS SUBMISSION FORM iii
DECLARATION OF ORIGINALITY iv
ACKNOWLEDGEMENTS v
ABSTRACT vi
TABLE OF CONTENTS vii
LIST OF FIGURES x
LIST OF TABLES xii
LIST OF ABBREVIATIONS xiii

CHAPTER 1 INTRODUCTION 1
1.1 Problem Statement and Motivation 4
1.2 Objectives 5
1.3 Project Scope and Direction 7
1.4 Contributions 7
1.5 Report Organization 7

CHAPTER 2 LITERATURE REVIEW 9

2.1 Previous works on business intelligence 9
2.1.1 Business Intelligence Systems 9
2.1.2 Business Intelligence Systems for Analysing University 10
Student Data
2.1.3 Business Intelligence for Educational Institution: A
11
Literature Review
2.2 Strength on the previous works 11
2.3 Limitation of the previous works 12
2.4 Proposed solution 13
2.5 Analysis and discussion 21
2.5.1 Overview of Power BI 22
2.5.2 Overview of Tableau 23

vii
 2.5.3 Overview of IBM Cognos Analytics 24
2.6 Feature of BI solution comparison table 25
2.7 BI feature pyramid 27
2.8 Business Intelligence Tools – Assessment Results 29

CHAPTER 3 SYSTEM METHODOLOGY/APPROACH (FOR 32

DEVELOPMENT-BASED PROJECT)
3.1 System Requirement 35
3.1.1 Hardware 35
3.1.2 Software 35
3.2 Activity Diagram 37
3.3 System Architecture 38

CHAPTER 4 SYSTEM DESIGN 39

4.1 System Block Diagram 39

CHAPTER 5 SYSTEM IMPLEMENTATION (FOR DEVELOPMENT- 40

BASED PROJECT)
5.1 Setting up 40
5.1.1 Software 40
5.2 Steps of data visualization and analysis 40
5.3 Timeline 41
5.4 Data collection and preparation 42
5.5 Visualization of results using Power BI 43
5.6 Visualize the existing results using Python 49

CHAPTER 6 SYSTEM EVALUATION AND DISCUSSION 56

6.1 Analysis and discussion 56

6.1.1 Analysis of dashboard 1 (Midterm) 56
6.1.2 Analysis of dashboard 2 (Practical) 58
6.1.3 Analysis of dashboard 3 (Assignment) 59
6.1.4 Analysis of dashboard 4 (Final score result) 60
6.1.5 Analysis of dashboard 5 (Final result) 60

viii
 6.2 Discussion of results 61
6.3 Recommendations and formulations 63

CHAPTER 7 CONCLUSION AND RECOMMENDATION 65

7.1 Project Review and Discussion 65
7.2 Novelties and Contributions 66
7.3 Future Work 67

REFERENCES 68
APPENDIX 72
WEEKLY LOG 102
POSTER 114
PLAGIARISM CHECK RESULT 115
FYP2 CHECKLIST 117

ix
 LIST OF FIGURES

Figure Number Title Page

Figure 2.1 BI system 9

Figure 2.2 Master Data and discovery 14
Figure 2.3 Self-service BI 15
Figure 2.4 Data Governance 16
Figure 2.5 Cloud BI 17
Figure 2.6 Real time analytics 18
Figure 2.7 Power BI 23
Figure 2.8 Tableau 24
Figure 2.9 IBM Cognos Analytics 25
Figure 2.10 The pyramid of BI tools features 28
Figure 2.11 Magic Quadrant for Analytics and Business Intelligence 30
Platforms
Figure 3.1 Flowchart of BI system 37
Figure 3.2 System architecture of data analysis 38
Figure 4.1 Block diagram of data analysis 39
Figure 5.1 Gantt Chart of Project 41
Figure 5.2 Step 1 of data visualization 43
Figure 5.3 Step 2 of data visualization 44
Figure 5.4 Step 3 of data visualization 44
Figure 5.5 Step 4 of data visualization 45
Figure 5.6 Dashboard overview of midterm result using Power BI 47
Figure 5.7 Dashboard overview of practical test result using Power BI 47
Figure 5.8 Dashboard overview of assignment result using Power BI 48
Figure 5.9 Dashboard of final score result using Power BI 48
Figure 5.10 Dashboard of final results using Power BI 49
Figure 5.11 Launch Jupyter Notebook in Anaconda Navigator 50
Figure 5.12 Locate the database saved 50
Figure 5.13 Data Cleaning 51

x
Figure 5.14 Select ‘Python 3 (ipykernel)’ 51
Figure 5.15 Read the data from the located file 52
Figure 5.17 Dashboard overview of midterm using Python 53
Figure 5.18 Dashboard overview of practical using Python 53
Figure 5.19 Dashboard overview of assignment using Python 54
Figure 5.20 Dashboard overview of final score using Python 54
Figure 5.21 Dashboard overview of final results using Python 55

xi
 LIST OF TABLES

Table Number Title Page

Table 2.1 BI tools chosen for comparison 22

Table 2.2 Business Intelligence tools and capabilities they offer 27
Table 3.1 CLO for UCCD9999 Website Design and Development 33
Table 3.2 Specifications of laptop 35
Table 3.3 System requirements for Power BI 35
Table 3.4 System requirements for Python 36
Table 5.1 Assessment Method and marks breakdown 43

xii
 LIST OF ABBREVIATIONS

BI Business Intelligence
CLO Course Learning Outcome
DW Data Warehouse
ETL Extraction, Transformation and Loading
OLAP On-line Data Processing
CRM Customer Relationship Management
DSS Decision Support Systems
GIS Geographic Information Systems
MD Master Data
DQM Data Quality Management
IoT Internet of Things
IA Information System Engineering
CS Computer Science
IB Business Information Systems
IT Information Technology
SQL Structured Query Language
NPL Natural Language Processing
IBM International Business Machine

xiii
Chapter 1

Chapter 1
Introduction

Cloud computing

Cloud computing has emerged as one of the most significant technologies in our current
technological era. As a result, cloud computing is defined as a computing sharing technology
that provides available computing resources, with cloud computing encompassing storage,
compute control, and application delivery via the internet as a service [5]. Aside from that,
Business Intelligence (BI) is characterized as a critical idea that has sparked widespread interest
in both academia and industry, with BI systems being used mostly in areas of business that
require decision-making to obtain value and insights. Business intelligence also refers to the
use of specialized tools and data to make decisions in a variety of organizational settings. Due
to the expanding relevance of BI systems, they are still the most widely and widely utilized IT
solution today. The overall goal of a business intelligence system is to help people make better
decisions.

Numerous studies have proven that higher education is critical to a country's economic progress
and prosperity. As a result, data is critical for major businesses, particularly higher education,
to make informed decisions. Because educational institutions have a high number of students,
employees, and activities, BI tools can be used to evaluate a vast amount of data to forecast
students' academic behavior, performance, and shortcomings. These statistics, for example, are
crucial throughout the decision-making process because they may be utilized to communicate
the need for change and improvement, which helps to improve the academic institution of a
particular institution's students [6]. BI solutions such as Power BI and Tableau give services
and assistance that aid top management in completely analyzing complex business
circumstances and streamlining decision-making processes.

With the availability of 'big data' and breakthroughs in machine intelligence, Business
Intelligence has become an exceptionally crucial term in this modern era of technological
growth [1]. BI systems had already generated significant attention in industry and academia
[2], and they've been extensively applied in many aspects of industry that require making
beneficial choices. Also, practitioners as well as academics required, however, get a better

1
Chapter 1

understanding of the mechanisms through which businesses might derives its value from BI in
addition to helping it reach its full potential. Studies that have investigated BI utilizing a variety
of hypotheses, investigation methods, and scientific data so far. Business intelligence (BI)
refers to the technologies, tools, and procedures that are used to collect, combine, evaluate, and
report business data. Aside from that, the basic goal of business intelligence is to allow people
to make better decision. Organizations are under mounting pressure to collect, analyze, and to
use their data to enhance decision-making as well as enhance company operations.

Business Intelligence

Business intelligence (BI) is a framework that enables a company to present the data, make
sense of the information, and knowledge into insight [3]. "A combination of techniques,
processes, systems, technologies, and tools that turn data into useful and valuable information
intended to provide greater strategic planning, tactical, as well as operational insights and
decision-making," as according to Forrester Research. Aside from that, there are also numerous
benefits of implementing business intelligence in our modern world as business intelligence is
more than just a software. One of the main benefits of business intelligence is faster analysis
alongside with intuitive dashboards. BI solutions are built to handle large amounts of data
processing in the cloud on the company's infrastructure. Also, BI tools collect data from many
sources and store it in a data warehouse, where it is then analyzed using user requests, drag-
and-drop analytics, and dashboards. In other words, Business intelligence dashboards make the
data analysis easier and simpler for non-technical people, allowing them to build narratives
with data without attempting to learn coding. It was also found that business intelligence
increased organizational efficiency. Leaders may use BI to access the information and get a
comprehensive perspective of their activities, as well as measure outcomes against the rest of
the company. Leaders may uncover regions of potential by taking a comprehensive picture of
the organization. As companies spend less time analyzing data and generating analyses, they
have much more opportunities to develop on new initiatives and services for the company [4].

Other than that, business intelligence provides better data driven business decisions as well as
improving employee satisfaction. IT staff and analysts devote fewer time to response to
requests from corporate users. Departments that previously couldn't view the personal data
without consulting analysts or IT may either do so with minimal training. BI is built to be

2
Chapter 1

scalable, so it can deliver data solutions to organizations that really need them as well as people
that desire it. Hence, business intelligence delivers trusted and governed data. Data
organization and analysis are aided by BI systems. Various departments' data is
compartmentalized in conventional data analysis, and consumers must visit many databases to
address specific information requirements. Sophisticated business intelligence solutions could
now merge every one of these internal databases with external sources like client data, social
data, as well as even weather data information into a single data warehouse. At the same time,
departments from across a company may view the existing data [4].

Cloud business intelligence

Cloud-based business intelligence, often also known as cloud BI, is defined as the process of
translating data into usable information in a cloud infrastructure, either partially or completely.
Without the cost or trouble of physical gear, cloud BI provides organizations with the
knowledge businesses ought to execute data-driven choices. Therefore, Cloud BI encompasses
all SaaS-based business intelligence [7]. In this advanced world, majority of business users and
consumers are adopting Cloud BI and there are several benefits to implement it. Presently, it is
very tough to locate a company which doesn't employ at least one cloud-based tool, such as
customer relationship management (CRM) software, online collaboration software, online data
storage, as well as certain help desk software. It's reasonable that businesses will want a cloud-
based bi solution as well. The transformation of data into valuable insights is at the heart of
business intelligence. It all boils down to having the appropriate information to the correct
individual time. The cloud is an excellent complement to this procedure. It allows you to
process, retain, send, and access that information in a quick and scalable manner [7].

It's crucial to understand the constraints of desktop BI solutions before considering the
advantages of cloud BI. If business users transfer data items from the cloud or client system to
additional devices via desktop BI, IT loses transparency and controls, which could also lead to
certification breaches and making version tracking challenging. With the help Cloud BI, all
these issues could be solved as well as providing other benefits too. One of the benefits is that
it is easy to deploy. Because cloud BI solutions are web-based, they are easy to set up and
doesn't require the help of IT specialists. Also, Cloud BI is mobile friendly, and it can be
accessed through mobile. Remote access is essential with the development of remote work.
With cloud BI, users are no longer restricted to on-site access and desktop PCs. Instead, its

3
Chapter 1

business intelligence (BI) products, such as Domo, can be accessed from any location or on
any device. Your customized dashboards are always available to you [7].

1.1 Problem Statement and Motivation

BI systems have been an important component of enterprises for decades, but there is a lack of
expertise and resources on how to successfully maintain these systems beyond the development
stage [25]. As a result, BI deployment is frequently beset by several complex processes,
obstacles, problems, risks, and downsides, making BI benefits difficult to realize. Integration
of data from several sources or systems into an organization's data warehouse, for example, is
a significant difficulty in BI implementation because it necessitates a large amount of effort.
Apart from the benefits of BI, academics and practitioners have discovered that BI systems are
costly, complex, resource-intensive, and difficult to manage. As a result, BI system deployment
involves more than just buying hardware and software; it is a complex project requiring proper
resources and infrastructure for a long period of time to stay up with management decision-
making [26]. According to some reports, BI systems use billions of dollars each year. Despite
this, more than half of these BI efforts are found to be ineffective [27]. [28] asserts that the
benefits of a BI system can only be realized if it is well deployed. Now, several studies have
revealed that understanding the factors for BI system implementation is critical, as they believe
and believe that the high rate of BI failure is due to a lack of understanding of the factors for
BI system implementation, and that these factors subsidies the realization of BI system benefits
[29].

Many firms struggle to manage their BI projects due to their complexity and ambiguity [30].
One of the problems with deploying a BI system, according to [31] is the organization's
readiness for BI adoption and implementation. Another challenge with BI deployment is the
lack of information in the organization regarding the benefits and prospects of BI systems, as
well as BI success criteria [32]. Furthermore, [32] stated that, despite the burgeoning market
for business intelligence and the complexity associated with its deployment, the beneficial
outcome determinants of BI system implementation remain poorly recognized. A few studies
on BI implementation were undertaken, and it was discovered that most BI system
implementation failures are due to a lack of understanding and preparedness to use BI systems
in many businesses [33].

4
Chapter 1

Another problem for BI deployment, according to [29], is how professionals evaluate them,
because experts have varied perspectives on BI. Experts in data warehouses, for example, may
regard BI as an add-on solution. Meanwhile, statisticians may consider BI to be a predictive
analysis-based tool, and data mining professionals may consider BI to be an advanced support
system that uses data mining techniques and algorithms. Organizations are hesitant to invest in
BI due to differing opinions of experts. As a result, there are various external and internal
aspects that influence the outcome of BI solutions, such as senior management support,
technical resource capabilities, the sorts of industries in which the firms compete, data source
quality, and investment funds [35]. Furthermore, integrating data from several sources or
systems into the organization's data warehouse is a key challenge in deploying BI because it
necessitates a great amount of effort [29]. As a result, managing the data warehouse is a difficult
task. Despite the many advantages of a business intelligence system, there are still several
issues with its deployment and use.

The aim of this thesis is to find out which components in a course where the students didn’t do
very well in terms of their CLO and cognitive C, practical P, and affective A as an indicator.
In this matter, students’ weaknesses in specific skills are analyzed to understand what are
causes of the underlying issue. At these times, many students from higher educational
institutions are dropping in terms of academic result. This project is essential due to vast
number of students who are starting to deteriorate in academic performance during in times of
pandemic. However, previous works on the existing project is not so accurate into analyzing
student performance as students’ weaknesses in certain skills or courses are remain unsolved
or lack of recommendations to overcome the issue. In this matter, it is crucial that this project
to be carried out to analyze deeply students’ academic performance with an improved method
to accurately find out which area of weaknesses categories are they into.

1.2 Objectives

First of all, the main objective of this project is to find out which area of weaknesses the
students based on the CLO and bloom taxonomy. In fact, the objective is also to identify the
problems related to students performance, which reported to have been deteriorating in this
current pandemic where all students are required to conduct their learnings at home remotely.
Students results are then analyzed by implementing the data analysis method to further propose

5
Chapter 1

a discussion result to find out what are their backgrounds or education which contributes to the
performance of students.

Secondly, the aim of the thesis is to study the software of business intelligence tools and to
carry out a comparison to seek for an appropriate tool for the project requirements. Due to high
availability of BI tools in the market, it is essential to find out which of these tools are suits the
needs of the project. To pick the right BI solutions on the huge amount of alternatives on the
current market, an exploration is required and find out which of the features are needed and
what are the features which are not necessary for the project’s needs. This comparison is vital
as BI tools come with many different features and complexity like tools with basic capabilities
and tools with advanced platform with more specialized features. Hence, the comparison will
enable us to find out which of the listed features should be used in which the solution is
adaptable easily, easy to handle, kind of support of functionalities being used, how good it can
handle security matters, as well as most vitally the cost of the solution. Therefore, after finding
out what is the right tool, the sample data of the students are collected and integrated into the
BI tool to visualize the student data by drag and drop queries to form a graphical representation
in a dashboard.

Third, the purpose of this research is also to analyse the results and outcome by collecting,
modelling, and analysing the data with the purpose to extract insights and information which
then provide a decision making to solve the matter. This is because when students’ data are
visualized in a form of graph, pie chart or bar chart, it makes the data easier to understand and
pull insights from by highlighting the trend and outliers as well as removing noise from data
and only focus on the important and useful information. This will enable us to find out what
are the weaknesses of students in particular field of a certain subject and propose a
recommendation to overcome the problem.

Lastly, the purpose of the research is to create a system of a business intelligence tool using
Python with integrated matplotlib to make use the business intelligence tools to carry out the
analysis. While using an existing tool to conduct the analysis, a simple system is developed to
automate and optimize the tiring process, create customized visualization as well as creating
machine learning modules and visualization of the predicted outcomes. In this case, the system
is created to enhance the data visualization as it helps us to tell stories by curating data in a
form which is easier to understand ad highlighting the trends as well as outliers.

6
Chapter 1

1.3 Project Scope and Direction

In this section, the project scope is to explore business intelligence tools which are available in
the market as a solution to overcome the issue of student academic performance as well as
creating a system alongside to do the analysis. This outcome of this project will enable us to
identify which area of skills or courses where students can’t perform well, and the CLO of the
students will be analysed, and a solution will be proposed to recommend suggestions that how
to overcome this issue. As due to the covid pandemic, many students are struggling to keep up
with their studies and studies have found out that students’ academic performance is
deteriorating and hence, the result of this project can guide us to monitor the performance of
each CLO as well as forwarding towards improvement in the course which is taught.

1.4 Contributions

The development of this project will enable lecturers or tutors from higher educational
institutions to gain better insight and understanding on students’ main area of weaknesses by
finding out if their factors of academic performance deterioration have a relation with their
educational background. Hence, the outcome of project will enable lecturers of tutors to know
what the knowledge or skills are required by end of the semester. Throughout this way,
lecturers may change the method or the way of teaching in class where it will be easier and
feasible way to convey vital information to students. In that case, more students will have a
huge leap in academic performance especially in these times of critical covid situation in our
country. In fact, the end of this project will reveal how much of most of the students could not
perform satisfyingly and the factors that contribute to such outcome are to be studied and
recommendations to be proposed to overcome this issue. Other than that, this project will show
the significance of applying both Power BI and Python in educational field and how it would
help in solving real world problems.

1.5 Report Organization

The details of this research are shown in the following chapters. In Chapter 1, the introduction
will briefly explain what business intelligence is as well as identifying the underlying issues
which is significant to solve. Therefore, the details in Chapter 1 will include the key details
regarding the project with the purpose of conveying the main message of the project scope.

7
Chapter 1

Secondly, Chapter 2 will conduct a literature review which enables us to find out what has been
done and what has not been done by other publishers in the field of business intelligence before
carrying out to propose a novel solution for the project. Furthermore, Chapter 3 will talk about
system methodology for the development based project where it will include methodologies
involved where it will be performed in several stages which are data collection and preparation,
Business Intelligence software selection and system development, data analysis and findings,
and then formulation of conclusions and recommendations. On chapter 4, it will further
discuss the structure of the business intelligence system design in the form of system block
diagram. This will provide us a better view of how the process of data visualization takes
place in several stages. On Chapter 5, system implementation will takes place where the
stages of data visualization will take place throughout the process. Aside from that, Python will
be used a programming tool to develop the existing result generated by Power BI with the
integrated platform, Jupyter Notebook with installed data analysis libraries like Matplotlib and
Pandas. On Chapter 6, the results of the data visualization are discussed analyzed based on the
graphical representation to find out which area of weaknesses the students are into, as well as
proposing a recommendation to solve the problems identified based on data analysis. In
Chapter 7, a conclusion is formed to summarize the results of the entire report and states the
novelties of this project, as well as the contribution of this project.

8
Chapter 2

Chapter 2
Literature Review

2.1 Previous works on business intelligence

2.1.1 Business Intellignce systems

Business intelligence (BI) is defined as the translation of information contained in knowledge

that enables us to provide appropriate information to a specific user at the right time to support
real-time decision making. As a result, BI brings together a variety of tools and technologies
that enable data collection, integration, analysis, and visualization.

To deploy a BI platform, various intermediate processes, such as the creation of a Data

Warehouse, must be completed, which is common in the development of this type of software
tool (DW) [15]. In this case, the Kimball process was used to design, develop, and deploy the
DW. As a result, several of these steps, such as task planning and expected results, describing
the architecture that will follow the BI system, selecting and installing the most appropriate BI
tool, building the Data Warehouse dimensional model, the Extraction, Transformation, and
Loading (ETL) process, and finally developing the BI application, are known to be critical for
the successful implementation of a BI system [15].

where: OLAP = on-line data processing, CRM=customer relationship management,

DSS= decision support systems, GIS = geographic information systems

Figure 2.1 BI system

9
Chapter 2

In addition, selecting the software that is most appropriate for attaining the intended results is
critical when developing a BI application. For example, it is vital to examine most of the
software on the market and select the kind that gives the required and desired resources.

2.1.2 Business Intelligence Systems for Analyzing University Student Data

The research work is presented in few stages starting from collection of data together with
preparation, selecting an appropriate Business Intelligence software and development of
system, analyzation of data and findings and then formulation of recommendations to tackle
the issue.

Throughout the first stage of the strategy, students' data is collected for data analysis. The data
comes from the lecturer's records who led exercise sessions in one of the University of National
and World Economy's university courses. The data is gathered using Excel software. The
dataset used for the study included 575 records of student performance in the course's exercise
classes over three academic years, from 2011 to 2013. During the exercise classes, the
researchers use two Microsoft Office software products: Excel (a software tool for creating
electronic tables) and Access (a software tool for creating relational databases), and they must
pass two assessments (on Excel and Access) over the course of the two semesters. The average
of the students' scores from the two tests is used to compute the student's overall score for the
exercise classes.

In the second step of the strategy, the software tools for developing Business Intelligence
applications are selected. Today's IT market offers a plethora of Business Intelligence tool
packages. Unfortunately, many of them are commercial items that are quite expensive and
accessing them for the purpose of research is normally prohibited. Fortunately, several of the
industry's largest IT manufacturers have already launched their own academic programmes and
are offering academic licenses to lecturers and professors who want to use the software for
research and education. QlikTech International was chosen as the business intelligence solution
due to its user-friendliness, ease of understanding and usage, and good graphical presentation
of the processed data and information. The developed QlikView application is divided into
three dashboards, each of which displays the study's most important data in the form of various

10
Chapter 2

graphical elements such as bar and pie charts, performance indicators, tables, and list boxes
(used as data filters on the various attributes).

The findings of the data analysis are presented and reported, conclusions are reached, and
recommendations are made for future research possibilities and data collecting during the
project's final step [35].

2.1.3 Business Intelligence for Educational Institution: A Literature Review

Research has also been done by managing organizations data to improve decision making. [36]
studies that certain relevant literature in the field of business intelligence in higher institution.
The methodology which was implemented were identification, selection, and feasibility. The
very first step were by searching through sites like IEEE, SpringerLink as well as Google
Scholar. On the other hand, the keywords which were used to search were ‘Business
Intelligence’, ‘Education in Business Intelligence’, ‘Higher institution business intelligence’.
After the previous is done, literature selection was performed by selecting based on title and
year of publication. The purpose of literature title selection is to identify whether it fits the field
of business intelligence in higher institutions or not. At the same time, a condition in a range
2000 to 2020 is limited for the selection of literature. Once the literature was identified and
selected, the leftover literatures are performed a review. Sustainability of according to research
were examined as well as the availability of literatures in complete text. [36] also stated that
chosen literature will be studied and there will be 14 articles to be discussed.

2.2 Strength on the previous works

First and foremost, one of the main strengths analyzed from the proposed previous work is that
they made good use with the adoption of business intelligence tool for analyzing variable data
could contribute to taking better and informed decisions. Other than that, the results of the data
analysis are well presented in detailed in different forms of graphical objects which includes
bar and pie charts, performance indicators, tables, and list boxes as well. Therefore, the decision
making is based on dashboards, scorecards, and reports, which allows graphical sharing of vital
data with the stakeholders. In this matter, a good BI dashboard allows you to monitor and
measure business performance and metrics. BI dashboard tools offer real-time centralized

11
Chapter 2

access to users enabling them to interact with and evaluate information, assisting them to make
smarter, data-driven decisions.

Hence, an appropriate business intelligence tool is chosen which is QlikTech where it brings
data and analytics together seamlessly with the only end-to-end, real-time analytics data
pipeline.

2.3 Limitation of the previous work

Based on my findings on the proposed previous work, it is found that there are several
limitations or weaknesses of implementation of BI tools in the education industry. One of the
very main issues is that CLO (Course Learning Outcomes) of students is not analyzed. CLO is
very crucial in terms of student academic performance as it observe or measure results that are
expected after a learning experience rather than seeing from academic results. While a student’s
result does not accurately indicate the understanding of student in certain courses, CLO must
be analyzed to gain insights of what the student have learnt throughout the semesters.

Furthermore, just a little amount of data is collected, resulting in low data quality. Due to the
restricted amount of data available, this could have an impact on decision-making, as data
correctness relates to how well it describes the real-world situations it is intended to describe.
Inaccurate data will cause problems because it will lead you to make the wrong judgement.
Because those findings and justifications are founded on erroneous data, the actions, and
decisions you make based on them may not have the desired results. For example, data may
lead a marketer to believe that the bulk of their clients are females in their twenties. Because
their consumers are largely males in their 40s, and their data is inadequate and erroneous, they
will wind up targeting the wrong demographic with their marketing, resulting in poor decision-
making.

In fact, I have also found out that the proposed work did not do a precise comparison of BI
tools available in the market. It is crucial to carry out an analysis of the comparison of business
intelligence tools to find out which are the necessary and which ones are the insignificant ones.
It is also important to know what exactly the organization needs or requires. By doing a
comparison, you will know BI tools will fits your requirements. Hence, it’s always vital that
the BI tools available in the market listed out alongside with its functionalities or features and

12
Chapter 2

carry out an analysis so that an appropriate tool will be chosen to do better decision making
effectively.

Finally, it is discovered that the proposed earlier work does not include python for data
analysis. Using Python and a business intelligence platform, we can generate stunning,
personalized visualizations, allowing us to create machine learning modules as well as
visualizations based on predicted outcomes. It can swiftly develop and manage data structures
thanks to the use of Python. Pandas, for example, offer a variety of tools for manipulating,
analyzing, and representing data structures and complex datasets. As a result, time series and
more complex data structures such as merging, pivoting, and slicing tables will be used to
develop new perspectives and views on current data sets.

2.4 Proposed solution

To begin, I propose that the problem be solved by selecting and executing the appropriate
business intelligence software tool. To pick the best BI tool among many available solutions
on the market, you must first determine what kind of features are necessary or compulsory
needed and which are not critical to the company’s demands. So, should we go with a basic
solution for everyday use or go for a more advanced platform with highly specialized and
advanced features? The solution is not direct, and towards the completion of this article, I will
expand on and justify a few of the options.

Second, most today's BI solutions (>90 percent) provide basic (standard) capabilities such
analyzation of data, ad-hoc reports, dashboards, visualization of data, performance metrics, ad-
hoc query, ad-hoc analysis, and key performance indicators (KPIs). When selecting a BI
solution, we must keep in mind that certain complex platforms with more advanced or
specialized capabilities may not have all the basic features. As a result, there are various
compromises between advanced functionality and the breadth of business intelligence tools.
It’s not unheard of for a product to make use of basic or advanced capabilities from other
solution providers.

First and foremost, I will offer a full inventory of advanced features and functionalities in
Business Intelligence systems, from which users should anticipate or demand the best solution
available. The following are some of the advanced features:

13
Chapter 2

• Master Data & Data Quality Management – To make accurate choices, solely
reliable data will be used. MD/importance DQMs are based on the demand for high-
quality, full, and accurate data. In this context, BI solution vendors have placed a strong
emphasis on master data management as well as data quality, and the quality of data
cycle. The data quality cycle encompasses every stages of offering high-quality data to
organizational users, including metric identification, evaluation, data mending as well
as cleaning, classification, archival, discovery, as well as evaluation [8]. Users could
track and data quality handled through a single, integrated source that contains the
master data, rather than by many, disjointed databases.

Figure 2.2 Master Data and discovery

• Data Visualization - Whenever it comes to data preparation for decision-making, BI

systems with extensive data discovery as well as visualization capabilities can help
eliminate the need for data scientists. The use of exploratory data discovery and
visualization together can help you overcome challenging data analysis barriers. Hence,
machine learning is also being merged into data discovery technologies in purpose for
assisting business analysts with every stages of the process, along-with devising to
investigation to presentation. Tools for integrating numerous sources, filtering,

14
Chapter 2

improving, and altering data to build novel data sets for use in analysis visualization or
up to date analytics should all be included in the solution.

• Self-Service BI – Consumers formulate and maintain reports and analyses within the
context of a well-tested and managed structure and tool set [9]. As according to Gartner,
business customers with self-service analytics capabilities will outperform experienced
data scientists. It will also be able to assist corporate users in learning ways to apply
and gain advantage from effective analytics as well as business intelligence
technologies, resulting in great business outcomes [9]. Self-service BI increases
efficiency and minimize the time to discovery, but it should not jeopardize the quality
or effectiveness of the outcomes. It must accomplish several objectives, namely
maintaining that greater speed does not threaten data trust and facilitating access to the
data and understanding. Improved data security and governance should assure data
accuracy and reliability, and the amendment should include all stakeholders and parties
responsible. So, it is crucial to fulfil architecture and requirements have been met while
maintaining a balance of flexibility and efficiency [10].

Self Service BI

Figure 2.3

• Data Governance – This technique guarantees that data satisfies specific criteria as
well as business needs when entering data into the database. Countless businesses have
endured massive losses in the recent because of a lack of data security [11]. Data
governance comprises designing a data framework that takes regulating policies and
procedures, evaluating, and maintaining data capital while involving people, stages, as

15
Chapter 2

well as technology into account to avoid these mistakes. Data governance that adheres
to regulations such as the Data Protection Regulation (GDPR) is also a great feature for
BI solutions [12].

Figure 2.4 Data Governance

• Cloud BI/Data Management - As stated by Gartner, numerous companies are going

to shift an important portion of its data undertakings to the cloud before 2021 [13]. As
a result, most Business Intelligence and data management vendors currently provide a
cloud-based service. Despite cloud Business Intelligence as well as data management
have much of the same functional capabilities as its on-premises versions, they usually
come at a lesser cost and put fewer load on IT departments

16
Chapter 2

Figure 2.5 Cloud BI

• Mobile BI - The advantages of letting employees or supervisors to make choices

regardless of where they are situated are considered by organizations. Because of its
volume to give powerful and precise data visualization within structure of graphs
including charts, dashboards, and performance measures, notepad and mobile devices
are improving effectiveness of mobile Business Intelligence, which had seen significant
commercial adoption [14].

• Deep Learning-Powered Analytics – Furthermore, Deep Learning is defined as a sort

of machine learning in which a neural model is trained into doing human-like activities
like speech recognition, responding natural language inquiries, and distinguishing
things and activities in pictures and videos. As per according to Gartner, deep learning
would be the dominant option strategy for data applications before 2023. [9].

• Real-Time Analytics – Data is evaluated using logic and mathematics to generate

insights that assist individuals in making improved major decisions. In certain
circumstances, real-time analytics implies that now the analysis has been conducted
within seconds or minutes of fresh data being received [9]. In several firms, faster data
reporting and analysis is a roadblock. To enable quicker and more fact-based
organization's decision, companies are increasingly depending on data from
transactional systems. Real-time analytics in BI may augment a company's existing
Business Intelligence strategy by offering novel, crucial data findings [15]. Enhanced
change data capture (CDC) technologies are being used by organizations to identify as

17
Chapter 2

well as gather data modification and data structures when they occur, as well as to alert
users of such changes.

Figure 2.6 Real time analytics

• Agile BI Development – Iterative development and information storage

commoditization are valued in this flexible and efficient system. This allows companies
to easily respond to different business requirements and minimizing total cost of
ownership. Agile BI involves fast prototype cooperation between business and IT,
allowing businesses to boost production speed whilst adapting to business requirements
more swiftly. The agile BI development process is complemented by project
administration, where it encompasses arrangement, requirements collecting,
development, and iterative functional, regression, as well as usability evaluation [15].

• Data Warehouse Modernization - New technical and business problems keep

businesses on their toes. They see the potential in using different data warehousing
architecture methodologies along with integrating extra technology options including
in-memory processing, cloud storage, including data storage automation instruments.
While competing with novel and less expensive implementation choices coming from
external service providers, IT must address evolving analytical requirements.
Multidisciplinary strategies must always be employed to suit the rise of requirements
of the market to optimize the business value of data. Therefore, it is time to assess how
new hardware and technology will improve and simplify corporate operations by
comparing obsolete data warehouses to present regulations [15]. Conventional data

18
Chapter 2

warehouses aren't developed to withstand the fast growth of data and the variety of data
kinds in which it generates, including big data. As a result, they aren't built to continue
alongside the changing needs of end users as well as the programmes which depend on
them.

• Data-Driven Culture - Modern manpower is so much more capable than ever before.
Therefore, the stages of education, engagement, as well as knowledge was at an all-
time high. A data-driven culture allows corporations to obtain every necessary data and
continue to exploit their values. Also, data-driven culture assists to elevate company
culture to the next level of performance by obtaining relevant KPIs rooted in business.
Also, transparency of data obtained KPIs is regarded as the part of aspect of the data-
driven civilization method of an organization [36].

• Data Preparation for Business Users - Data cleaning, structuring, and enriching are
called the stages of cleaning, structuring, and enriching data. These are the steps for
exploratory and advanced analytics. Hence, preparation of data is all about giving
people the tools they need to make their data satisfy their agenda without having to use
IT. Technology that is easy for people to use and understand, as well as smart coaching
and quick outcomes, were needed with the purpose to get data preparation to business
people [15].

• Integrated Platforms for BI/Performance Management - Many businesses want to

use the same tool for both BI and project management. The trend has led to one of the
most common and long-term things in the business intelligence market, as well [15].

• Embedded BI and Analytics – In operational applications, intelligence is being

applied. Embedded business intelligence and analytics refers to the incorporation of
business intelligence software aspects (such as dashboard outlining, visualizing data, as
well as analytics instruments) into non-business intelligence applications. Customers
gain from a far more pleasant and user-friendly user experience when using integrated
business intelligence, which is also its key benefit over solutions which require two
distinct platforms [16].

19
Chapter 2

• Data Storytelling – Data visualizations, infographics, dashboards, data presentations,

and a range of other tasks are all handled by data storytelling. It's not only about making
aesthetically appealing graphs. Data storytelling is described as a method of expressing
data insights that incorporates three key components: data, visuals, as well as narrative.
By integrating the proper visuals and narrative alongside the correct information, a data
story which able to affect as well as drive change may be generated [17].

• Using External/Open Data - Consumer, market, climatic, regional, and demographic

data, as well as similar analytical findings, may all give significant information. These
and many other kinds of data may be provided to firms by other Business Intelligence
generalists, specialist service providers, or data exchange platforms. Also, open data is
always utilized to create models that are based on focused research [18].

• Analytics Teams/Data Labs – Described as independent business segments developed

specifically to get a company's data science programme off the ground. They'll need to
invest in novel technology to keep, process, as well analyses information. When
analytics improves, so does the importance of solution implementation and efficiency.
It's a novel stumbling barrier for software solution providers, and then it involves the
creation of new organization structure to integrate data labs, IT departments, and
business divisions [19].

• Visual Design Standards – Described as a method of presenting significant data in a

manner that is both effective and efficient. The tendency of visual design qualities has
evolved in the past three years because of the rising requirement to analyses massive
volume of data to remain competitive and give its most direct results. Support for visual
design standards is seen as a need that BI vendors must meet in software selection
procedures [19].

• IoT Analytics - Because as internet of things (IoT) advances with novel sensor, device,
as well as wireless technologies, significant economic value must be identified in
analytics instead than new hardware. Sellers are beginning to provide these services to
their customers, as well as extending their service portfolios into new business sectors.
IoT data necessitates real-time data processing. In addition, the variety of IoT data

20
Chapter 2

mandates the use of new architecture, tools, and procedures to effectively process,
retain, as well as evaluate IoT data [37].

• Big Data Analytics – gives the capacity to examine large, varied, and velocity data sets
obtained from different sources, such as textual, sensory, geographical, and clickstream
data. Inside these arrangements, big data analytics should be able to analyses massive
datasets in real-time or near real-time, which includes modelling, visualization,
forecast, and improvement. Businesses are turning to big data analytics to help them
make better decisions and improve their processes [20]. Structured information
analytics, sentiment analysis, web analytics, audiovisual analytics, social network
business intelligence, and systems analysis are all examples of big data analytics.

• Data Lake – Based on James Dixon's original proposal, a huge storage of data in its
inherent state of unstructured and structured data [21]. Without any purification,
standardization, remodeling, or modification, it is not feasible to keep and analyses data
in its raw and original form, direct from the data sources. Data lakes enable for ad hoc
queries, data analysis, and findings analytic searches since information management
and architecture might well be adjusted on the fly during runtime. [38]. The goal of a
data lake is to keep all information on the same system, encompassing relational, non-
relational, including big data. [22].

• Edge Computing and NLP Its purpose is to make NLP computation more responsive to
user demands. BI is paying attention to efforts to understand user behavior, attitudes,
and emotions [23]. The usage of NLP in combination with DL is required to
comprehend verbal and nonverbal messages. AI has made significant progress in this
direction recently, and even more BI solutions that include speech or language
understanding technologies are expected to develop soon [24].

2.5 Analysis and discussion

After listing the features of BI tools, we'll conduct an examination of open-source business
intelligence solutions available in the market and compare those features as well as

21
Chapter 2

functionality. We've whittled down the number of business intelligence products to three.
Using methodology to enable organized study and comparison of BI solutions will provide
greater insights to the present status of industry as well as assist in the selection of a business
intelligence tool that includes all the aspects. Furthermore, the research conducted allows us to
detect current trends in BI solution development. This study can be used as a road map for
selecting a business intelligence solution that meets the needs of the enterprise. As a result, we
produced a feature-to-BI solution cross-comparison table and a feature pyramid to reflect the
latest trends in BI solution development, with a focus on unique and future features.

Table 1 shows three different BI tools which are currently on the market, alongside a link to
each of the descriptions. Most of these tools are currently the most popular business intelligence
solutions. As a result, according to user reviews, these solutions are the top business
intelligence and analytics software of all time. Essentially, the list comprises simple-to-use
systems that enable a wide range of analytic work capabilities without requiring much IT
participation, allowing users to quickly embrace BI tools. For example, tools can pre-define
data models as a requirement to analysis, and in some situations, they can even generate a
reusable data model automatically. As a result, we will assess the functionality of the BI tools
listed in Table 1 by comparing them to the proposed basic and advanced features of BI tools in
the next section.

Business Intelligence tools Web Page

Power BI https://powerbi.microsoft.com/en-us/
Tableau https://www.tableau.com/
IBM Cognos Analytics https://www.ibm.com/products

Table 2.1: BI tools chosen for the comparison

2.5.1 Overview of Power BI

Microsoft Power BI is a set of tools, software services, and connectors that work together to
transform disparate data into visually appealing and interactive insights. Power BI can operate

22
Chapter 2

with both simple data sources like as Microsoft Excel and more complex ones such as cloud-
based or on-premises hybrid data warehouses. Power BI also allows you to connect to your
data sources, visualize, share, and publish your findings to whomever you want.

Besides, Power BI is simple and fast enough to connect to an Excel workbook or a local
database. Hence, it can also be robust and enterprise-grade in which it’s ready for extensive
modeling as well as real time analytics. This means it can be implemented in various
environments from a personal report and visualization tool to the analytics and decision engine
behind group projects, divisions, or entire corporations.

Due to Power BI is a Microsoft product and has built in connections to Excel, there are lots of
functions that will be familiar to an Excel user.

Figure 2.7 Power BI

2.5.2 Overview of Tableau

Tableau was established in 2003 because of a computer science project at Stanford that aimed
to enhance the flow of analysis and allow the data more accessible to people through
visualization. Co-founders Chris Stolte, Pat Hanrahan, and Christian Chabot developed and
patented Tableau's foundational technology, VizQL where it is visually expresses data by
translating drag-and-drop actions into data queries via an intuitive interface. Since their
foundation, they have continuously invested in research and development at an unrivaled pace,
developing solutions to assist anyone working with data to receive faster and find out
unanticipated insights.

23
Chapter 2

Furthermore, making machine learning, statistics, natural language, and smart data prep more
usable and dependable to assist human ingenuity in analysis is part of this. They also offer not
only a comprehensive, integrated analytics platform, but also proven enablement resources to
help companies implement and scale a data-driven culture that delivers resilience and value
through compelling outcomes. In 2019, Tableau was acquired by Salesforce, but its aim
remains the same to help people see and understand their data. Currently, companies of all
sizes, from non-profits to multinational corporations, and from every industry and department
are using Tableau to empower and encourage their employees to make data-driven decisions.

Figure 2.8 Tableau

2.5.3 Overview of IBM Cognos Analytics

IBM, a web-based reporting, and analytic tool is designated as Cognos Business Intelligence.
Its goal is to do data aggregation and generate thorough reports that are easy to understand.
Graphs, several pages, different tabs, and interactive prompts can all be stored in reports. Web
browsers or mobile devices such as tablets and smartphones can be used to access these reports.

Aside from that, Cognos gives you the option of exporting the report in XML or PDF format
or viewing the report in XML format. As a result, you may schedule the report to run in the
background at a specific time, which saves time in terms of monitoring the report on a regular
basis because you don't have to run it all the time.

Also, IBM Cognos has a wide range of capabilities and may be considered enterprise software
because it provides a versatile reporting environment that is suitable for both large and small

24
Chapter 2

businesses. Power Users, Analysts, Business Managers, and Company Executives will find it
useful. Analysts and power users want to be able to produce ad hoc reports and different views
of current data. Business executives demand dashboard layouts, cross tabs, and visualizations
to examine summarized data. Cognos makes both options available to all users.

Figure 2.9 IBM Cognos Analytics

2.6 Feature to BI solution comparison table

At this part, 3 business intelligence tools are evaluated against their functionalities using the
product/functionality matrix produced in Table 2 to create a product/feature matrix. As a result,
the green checkmarks will show tools that are integrated, while the red 'x' will represent tools
that are lacking. Furthermore, these aspects enable us to examine the features of a given BI
solution and compare them to those of other systems. In this case, it will provide us an
understanding of the present state of the BI market and assist us in identifying solutions that
make use of advanced features.

Based on the analysis, it is revealed that over 90% of business intelligence tools provides data
analysis, ad hoc analysis, dashboards, ad hoc query tools, KPI’s, and performance metrics that
belong to the basic group of features. Other features include data management, data
exploration, self-service BI, saas BI, mobile BI, deep learning, and big data analytics. It's
important to note that, while most suppliers offer the entire range of capabilities in their
products, they isn't quite as practical or high-quality as capabilities found in other solutions.

25
Chapter 2

The research also revealed that most solutions allow for the merging of certain functionalities
by other companies. Advanced visualization capabilities of one solution, for example, able to
be combined to different solutions, purpose for delivering the correct visualization solution to
their consumers. Instead of developing their own solutions, sellers are expected to integrate the
greatest-rated solutions from other providers to seller’s products.

Functionalities Power BI Tableau IBM Cognos My

proposed
solution
Master Data ✔ ✔ ✔ ✔

Data Discovery ✔ ✔ ✔ ✔

Self-service BI ✔ ✔ ✔ ✔

Data ✔ ✔ ✔ ✔
governance
Cloud BI ✔ ✔ ✔ ✔

Mobile BI ✔ ✔ ✔

Real Time ✔ ✔ ✔ ✔
Analytics
Agile BI ✔ ✔ ✔ ✔
development
Data ✔ ✔ ✔ ✔
warehouse
Data driven ✔ ✔ ✔ ✔
culture
Data ✔ ✔ ✔ ✔
preparation for
business users
Integrated ✔ ✔ ✔ ✔
platforms for
BI

26
Chapter 2

Apps ✔ ✔ ✔ ✔
Integration
Embedded BI ✔ ✔ ✔ ✔
and analytics
Data ✔ ✔ ✔ ✔
Storytelling
Using ✔ ✔ ✔ ✔
external/open
data
Analytics team ✔ ✔ ✔ ✔

Visual design ✔ ✔ ✔ ✔
standards
Big data ✔ ✔ ✔ ✔
analytics
Data Lake ✔ X ✔ ✔

IoT analytics ✔ X ✔ ✔

Augmented ✔ X ✔ ✔
analytics
Deep learning ✔ X ✔ ✔
powered
Edge X X X X
computing and
NLP

Table 2.2: Business Intelligence solutions and the capabilities they offer

2.7 BI feature pyramid

In this section, we'll go over the feature pyramid, which demonstrates BI functionalities via
those frequencies with which they appear in BI solutions (see Figure 2.7). The most
fundamental features, which are implemented in most Business Intelligence tools, are placed
at the base of the pyramid, whereas rare functionalities are at the highest peak.

27
Chapter 2

The base layer, first and foremost, embraces elements that we classed as fundamental, and
which are found in 60 percent of BI systems today. The second tier includes augmented
analytics, IoT analytics, as well as data lakes, where they were employed in around a quarter
of business intelligence solutions. Also, nearly 40% of the evaluated solutions said
organizations had began using augmented analytics in a more advanced form. Solutions gives
business decision makers innovative visual access to complex data, backed by clever
recommendation algorithms that help them comprehend hidden and useful insights. As a result
of the use of automation and embedded intelligence, the time required for data preparation has
been drastically reduced. The investigation, however, reveals that every of the solutions are
strongly focused on augmented analytics. Meanwhile, deep learning-powered analytics make
up the third layer. The report shows that about a quarter of the analyzed Business Intelligence
solutions provides alternatives of deep learning analytics.

Finally, developing capabilities such as edge computing and natural language processing
(NLP) enable Business Intelligence solutions to acknowledge human perceptions, behavior, as
well as feelings. Even though none of the systems we examined give such possibilities (0%).
However, based on current patterns, we can confidently predict that these features will be
included in future BI editions.

Figure 2.10 The pyramid of BI tools features

28
Chapter 2

2.8 Business Intelligence Tools – Assessment Results

After analyzing the business intelligence tools based on Table 2, it could be concluded that the
most appropriate tool to analyze student performance would Power BI. Throughout my
findings, I found out that Power BI is an easy-to-use tool for data analysis and visualization
due to the user-friendly interface unlike other tools which adopts highly complex layouts. Due
to this matter, it’s safe to say that this tool is a perfect starter choice for beginners who has little
or no experience in data analytics. In fact, this tool performs at great speed and better in terms
of performance when the data volume is limited. While this tool tends to drag slowly while
handling large data, it fits anyone who just need it analyze small amount of data. As for my
project, no bulky data is required, therefore this tool suits perfectly for simple visualization.
Comparing to Tableau which able to handle bulky volumes of data, at the same time it has high
complexity layout and normally made for professionals to use. In this case, it wouldn’t be a
beginner friendly tool for starters who begin to adopt business intelligence in the market.

As previously said, Power BI's user interface is easy, allowing it to readily interact with other
Microsoft products such as MS Access, MS Excel, and others. The Power BI interface is simple
to understand and grasp, even for people with no prior expertise or experience in the field of
business intelligence, and it is user-friendly, allowing you to use more efficiently. The report,
data and the model view are all available on the left side of the canvas in Power BI desktop.

Furthermore, compared to Tableau, Power BI offers a variety of data sources but has limited
access to other databases and servers. Microsoft Excel, Text/CSV, Folders, MS SQL Server,
Access DB, Oracle Database, IBM DB2, MySQL database, PostgreSQL database, and others
are among the data sources. Because the only database I needed for my project was a MySQL
data source, this programme will provide enough capabilities and functionality to complete my
project. When it comes to simplicity of use, Power BI has an advantage because its user
interface is built on Microsoft 365, which most of the users are familiarize with.

Furthermore, Power BI has a drag-and-drop capability that is simple to utilize. It has features
that make data more aesthetically appealing. It has many detailed and appealing visualizations
that can be used to generate reports and dashboards. You may ask questions about your data
with Power BI, and it will provide you with useful insights into where the benefits are found.
Power BI, on the other hand, has the advantages of Microsoft Business Analytics, which
includes platforms like Azure Machine Learning, SQL Server-based Analysis Services, real-

29
Chapter 2

time data streaming, and many Azure databases. This aids in the comprehension of data as well
as the analysis of trends and patterns in the data.

Figure 2.11 Magic Quadrant for Analytics and Business Intelligence Platforms

Although Power BI is still new in the business intelligence field, the magic quadrant shows that
it is quite competitive in the market when compared to Tableau or IBM Cognos Analytics.
Power BI is, in fact, frequently used by corporate executives and organizations. Within the
Leaders' quadrant, Microsoft Power BI is positioned furthest to the right in terms of
completeness of vision and furthest up in terms of capacity to execute, according to my study.
One of the reasons is that it is cloud mature, relying on one of the fastest growing BI clouds.
In fact, Power BI Mobile is without a doubt the best-in-class mobile experience, allowing end-
users to easily connect to and interact with both cloud and on-premises data, and take it with
them wherever they go.

To summaries, all the standards listed in Table 2 were chosen with the education industry in
mind. As a result, Power BI was chosen to construct a BI application and demonstrate the
efficacy of Business Intelligence (BI) tools in analyzing student performance, based on these
requirements. Power BI Suite covers all the standards and offers more benefits than other BI
products like as Tableau and IBM Cognos Analytics. Furthermore, it is a tool with a good and
better interface, as well as a range of distinct representations. However, we must keep in mind

30
Chapter 2

that the tools discussed and reviewed in this section are continually expanding and changing,
and functionality that they currently lack may become available soon.

31
Chapter 3

Chapter 3
System Methodology

Proposed Method/Approach

This section will include the methodologies involved where it will be performed in several
stages which are data collection and preparation, Business Intelligence software selection and
system development, data analysis and findings, and then formulation of conclusions and
recommendations.

During the first stage, data collection and preparation, the sample data will be created
and prepared for analysis. In this phase, it starts off with the identification of the research
domain. UCCD9999 Website Design and Development methodology module was picked for
the research domain and the course learning outcomes (CLO) for this module is studied. In
short, the course has an objective to teach students regarding the expert system development
methodology by utilizing an expert system development life cycle. Therefore, the
developments of the CLO for this course are based on the Bloom Taxonomy which is as shown
in Table 3. According to the Bloom taxonomy, the three domains are consisting of cognitive
domain, affective domain, and psychomotor domain in which they are applied to CLOs of the
course. Thus, a few evaluation methods are implemented to test the students’ understanding on
the materials taught throughout the semester. The assessments are consisting of midterm
(20%), assignment (20%), practical test (10%) and final examination (50%). Aside from that,
this subject will give a brief introduction to the constructing web application which is
progressive as well as working web application with the implementation of knowledge and
skills in HTML, JavaScript, Bootstrap, CSS, and event handing programming. The table below
demonstrates the CLO of students at the end of the semester.

32
Chapter 3

CLO Course Learning Bloom taxonomy

Outcome
CLO 1 Create static, dynamic, and Cognitive domain
interactive web pages
using
different technologies
CLO 2 Apply knowledge of event- Cognitive domain
driven and object based
programming concepts in
client-side web
development
CLO 3 Apply emerging front-end Cognitive domain
web technologies in the
design
and development of web
applications
CLO 4 Able to develop responsive Psychomotor domain
as well as progressive web
applications

Table 3.1 CLO for UCCD9999 Website Design and Development

The software which will be used to store the database as data collection is Microsoft Excel.
The dataset will include a total 239 number of students who will be enrolling in this course. In
this stage, the marks of the students are collected based on their respective coursework
(midterm, practical test, and assignment) totaled up 50% and a final exam which is 50%. The
following results of students will be imported into Power BI to undergo interactive data
visualization and analytics. By doing this, it allows us to visualize data into meaningful insights
to explore solutions to overcome the issue faced which to find out what area of weaknesses of
students in. Aside from that, alternative software, Python will be implemented as well to
visualize data into graphical presentations with pre-installed matplotlib and pandas.

33
Chapter 3

During the second stage, software tool is selected for the Business Intelligence to analyze the
student performance. Nowadays, there are lots of Business Intelligence tools available in the
IT market. Unfortunately, most of them are very costly and very limited access for research
purposes. Luckily, there is an alternative solution where Power BI will be chosen as the
appropriate tool for this project. It is very easy to use and navigate and it packs with limited
features which are enough to utilize for basic usage. Similarly, to other BI tools, it offers data
preparation and discovery, amazing and interactive dashboards, and rich visualizations and
most importantly it is free for the basic features. Due to its simplicity, the data can be visualized
in just a few clicks, and it has a user-friendly interface. Currently, Power BI is one of the best
tools available on the market which adopted by many business leaders around the world and
widely used by people. Based on the above-mentioned reasons, it is concluded that Power BI
will be used a business intelligence tool. Hence, a simple system will be developed as well to
perform the data analysis other than using the existing tool. In this development of the system,
Python will be top pick for data analysis due to its simplicity and readability and at the same
time offering plethora of useful options for data analyst. Hence, it is a very good programming
language for novice or beginners to start off with, where you can use it its relatively easy syntax
to form effective solutions for complex situations by using a few lines of codes. Therefore,
Python will integrate with Power BI to write codes of script to visualize the graphical
visualization on the BI tool.

In the next stage, data analysis and findings, Power BI and Python will be used as the business
intelligence tool to conduct the visualization of student results. Anaconda is downloaded to use
Jupiter Notebook as it has integrated Python 3.6 with pre-installed matplotlib. The results of
the students are analyzed and organized in several dashboards in which the most essential
information from the data analysis is shown in the form of different graphical objects such as
bar and pie charts, performance indicators, tables as well as list boxes.

In the last stage, formulations and recommendations, the results and outcomes of the data
analysis are shown and presented and then described based on the finding. Weaknesses of
students in the course are identified and recommendations are drawn to overcome the issue
which is to be proposed and conclusions are drawn as well for future research opportunities
and ad data collection.

34
Chapter 3

3.1 System Requirement

3.1.1 Hardware

The hardware which will be involved in this project will be a computer. A computer will be
issued to carry out the data analysis and to perform data visualization with Power BI and
Python. The specifications of the laptop needed to carry out the project is as shown as below:

Description Specifications
Model Asus A510U series
Processor Intel Core i5-8250U
Operating System Windows 10
Graphic NVIDIA GeForce GT 940MX
Memory 12GB DDR4 RAM
Storage 1TB SATA HDD

Table 3.2 Specifications of laptop

3.1.2 Software

The software which will be involved in this project will be Power BI. The business intelligence
tool is issued to run reports and surface insights based on a data collected. Therefore, the
requirements of the software needed to carry out this project is as shown as below:

Description Requirements
Operating system Windows 8.1 / Windows Server 2012 R2, or later
Memory (RAM) At least 2 GB available, 4 GB or more recommended
Display At least 1440x900 or 1600x900 (16:9) required
CPU 1 gigahertz (GHz) 64-bit (x64) processor or better recommended

Table 3.3 System Requirements of Power BI

35
Chapter 3

The next software which will be included in this project will be Python. The programming
software is issued to manipulate, analyze, and represent data structures and complex dataset as
well as visualizing the graphical objects. The requirements of the software to perform the tasks
is as shown in below:

Description Requirements
Operating system Windows 7 or 10
Memory (RAM) At least 4gb ram
Storage At least 5GB disk free
CPU Minimum Intel Atom® processor or Intel® Core™ i3 processor

Table 3.4 System Requirements of Python

36
Chapter 3

3.2 Activity Diagram

Figure 3.1 Flowchart of BI system

37
Chapter 3

3.3 System Architecture

Figure 3.2 System architecture of data analysis

38
Chapter 4

Chapter 4
System Design

4.1 System Block Diagram

Figure 4.1 Block diagram of data analysis

Firstly, a group of students’ data are collected based on their respective scores on midterm,
assignment, practical and final exam. The group of students consist of different programmes,
Business Information Systems, Information Systems Engineering and Computer Science. The
next process will be data cleaning, where it’s a process of fixing, incorrect, incomplete or any
null numbers within the dataset collected. Aside from that, this is also to ensure that only the
most significant data is filtered to ensure accuracy and reliability of data before conducting
data analysis. The next step will be visualizing the results by using an open source business
intelligence tool which is chosen from the available market. The data is then transferred and
integrated into Power BI and the graphical representations are then produced. The existing
results are well as developed using Python with installed libraries of matplotlib and pandas for
data visualization. Both of the reports are then generated with all the stats of the students’
academic performance in the form of different graphical representations. The reports are then
analyzed and discussed which then further given recommendations for decision making.

39
Chapter 5

Chapter 5
System Implementation (For Development-Based
Project

The goals of preliminary data analysis are to modify the data to make it ready for subsequent
study, define the data's essential properties, and summarise the findings. The theoretical and
practical techniques to achieving the desired objectives are covered in this chapter.

5.1 Setting up

5.1.1 Software

Before we begin visualizing the data, there are three software needed to be installed and
downloaded in my laptop:

1. Power BI
2. Python 3.9
3. Anaconda navigator (with matplotlib installed)

5.2 Steps of data visualization and analysis

The hardware involved in this project is computer alongside with a few software to do the data
visualization. A sample data of 239 students with different course background are transferred
into raw data before integrating the data into business intelligence software to graphically
present the visualization in the form of dashboards as well as visualizing the data using a
programming software, Python. Furthermore, an analysis will be conducted based on the results
processed.

The following are the steps of data visualization process:

i. Data collection and preparation
ii. Visualize the results using Power BI
iii. Visualize the existing results using Python
iv. The graphical presentations are analyzed and discussed

40
Chapter 5

v. Recommendations and formulation to improve students’ performance

5.3 Timeline

To organize the tasks in a more systematic way, a gantt chart is created to plan the sub activities
which will be useful for planning and scheduling Project 2 for this semester. The gantt below
demonstrates what need to be done in a precedence order to avoid confusion as the each task
is contingent on the completion of another. Based on the gantt chart below, 12 weeks are given
to complete Project 2 where each task shall be done by week 1 till the end of completion. The
starting date of the project will be on Week 1, 13 June 2022 till the end of the project ending
at Week 12, 29 August 2022.

Figure 5.1 Gantt Chart of Project

41
 Chapter 5

5.4 Data collection and preparation

In this process, we collected a handful of students’ marks based on each of their respective
course, UCCD9999 Website Design and Development where the coursework will be divided
into three parts which are midterm test, practical test, and group assignment and a final exam.
The students who took this course consists of different students from different programmes of
study which are CS (Computer Science), IB (Business Information Systems), and IA
(Information Systems Engineering). Therefore, Table 2 summarises the percentages
distribution of each question according to CLO based on the categorisation as shown in below.

N Program Bloo Delive Assessment method and marks breakdown

o me m ring
learning Tax Metho Midt Practical Assignm Final Final Final Final
outcome ono ds erm test ent exam exam exam Exam
(PLO) my Q1 Q2 Q3 Q4
20% 10% 20% 12.5 12.5% 12.5% 12.5%
%
1 CLO1 Cog Lectur
nitiv e,
/ /
e Practic
al
2 CLO2 Cog Lectur
nitiv e,
/ /
e Practic
al
3 CLO3 Cog Lectur
nitiv e,
/ /
e Practic
al

42
 Chapter 5

4 CLO4 Psyc Lectur

hom e,
/
otor Practic
al

Table 5.1 Assessment Method and marks breakdown

5.5 Visualization of results using Power BI

In this process, we will be implementing the software, Power BI as our business intelligence
tool to visualize the output of the students’ results. In this first step, the sample data of students’
mark which in the form of Excel format are integrated into Power BI. Therefore, to begin, we
get data from the Excel workbook to locate our database which to be integrated into the
software.

Figure 5.2 Step 1 of data visualization

43
Chapter 5

In our next step, before we load the data, we would transform the data first to do some
modification to the queries like renaming the column names, removing null fields, and filter
out only the significant information we need for analysis. This is done so that we could
consolidate them into a useful and meaningful query.

Figure 5.3 Step 2 of visualization

In our third step, we had renamed and filtered the rows and it is ready to be loaded into the
platform by clicking “Close & Apply” to save the changes made to the queries.

44
Chapter 5

Figure 5.4 Step 3 of visualization

In our final step, we visualize the data in the form of graphical representations like stacked
column chart, pie-chart, table, funnel, clustered bar chart, area chart, Tree map, and line chart
by dragging the data on the right side into the visualization fields. The first dashboard, which
presented in Figure 4.5 provides the brief information of midterm results according to average
scores, programme, gender, and midterm results. and student id.

Figure 5.5 Step 4 of visualization

The Power BI application consists of five dashboards where the first dashboard, presented in
Figure 4.5 conveys the overall information regarding the midterm test results of students
according to average marks of midterm by all students, average marks by programme, average
marks by gender, average marks by programme and gender, number of students who
pass(marks =>50), number of students who failed(marks < 50), students gender who take the
course and programme of students who took the midterm.

45
Chapter 5

The second dashboard provides the information average of practical test by student and
programme, average by programme, average by gender, maximum of student marks for each
programme, median of student marks by programme, number of students who pass (marks
=>50) and fail (marks <50), maximum by programme and maximum practical test by gender.

The third dashboard provides the information average of assignment by student, average by
programme, average by gender, maximum of student marks for each programme, median of
assignment by programme, number of students who pass (marks =>50), maximum by
programme, minimum assignment by programme and maximum assignment by gender.

The fourth dashboard provides the information of final score by students, average of finals
score by programme, the number of students who passed(=>40), average of finals score by
gender, average of final exam Q1, Q2, Q3, Q4, average of final exam Q1 by programme,
average of final exam Q2 by programme, average of final exam Q3 by programme, and average
of final exam Q4 by programme.

The fifth dashboard provides the information final result(100%) by students, average of final
exam by programme, number of students who passed(=>50), number of students who
failed(50<), average of final result by gender, max of final result by programme, median of
final result by programme, and average of final exam and coursework.

46
Chapter 5

Figure 5.6 Dashboard overview of midterm result using Power BI

Figure 5.7 Dashboard overview of practical test result using Power BI

47
Chapter 5

Figure 5.8 Dashboard overview of assignment result using Power BI

Figure 5.9 Dashboard of final score result using Power BI

48
Chapter 5

Figure 5.10 Dashboard of final results using Power BI

5.6 Visualize the existing results using Python

In this process, similar results of Power BI output are achieved with the use of Python with
integrated matplotlib and panda. In this matter, Anaconda navigator is downloaded into system
which enables us to launch application as well as manage conda packages, environments, and
channels without the use of command line commands. Once the application is successfully
installed, launch the Jupyter Notebook which is an open-source web application for the purpose
of scientific computation, data visualizations and data processing. In short, this web application
is a useful tool for writing as well as iterating on Python data analysis code. With the usage of
this tool, you can create lines of code and execute them at the same time instead of creating
and rewriting a full programme.

49
Chapter 5

Figure 5.11 Launch Jupyter Notebook in Anaconda Navigator

In our next step, locate the database saved in the folder in Desktop

Figure 5.12 Locate the database saved

50
Chapter 5

Modify the data source by data cleaning so that it is free of errors and anomalies as clean and
consistent data allows us to perform data visualization easier.

Figure 5.13 Data Cleaning

In our fourth step, click New and select ‘Python 3 (ipykernel)’ to start visualizing data.

Figure 5.14 Select ‘Python 3(ipykernel)’

Next, import pandas and matplotlib library with the purpose to create static, animated, and
interactive visualizations. Then transport the sample data from located file into the web
application and display the data which needed to be visualize later.

51
Chapter 5

Figure 5.15 Read the data from the located file

Visualize the data with the implementation of matplotlib and panda’s library to produce the
clustered bar chart with sorted descending order. The bar chart below is visualized based on
the average of midterm(y-axis) and student ID(x-axis) alongside with title and legend.

Figure 5.16 A bar graph output of midterm

52
Chapter 5

Similar results are achieved with the use of Python with integrated matplotlib and pandas. The
first and second dashboard is shown below as the data analysis of midterm and practical result
of students for the course UCCD9999.

Figure 5.17 Dashboard overview of midterm using Python

Figure 5.18 Dashboard overview of practical using Python

53
Chapter 5

Figure 5.19 Dashboard overview of assignment using Python

Figure 5.20 Dashboard overview of final score using Python

54
Chapter 5

Figure 5.21 Dashboard overview of final results using Python

55
Chapter 6

Chapter 6
System Evaluation and Discussion

6.1 Analysis and discussion

6.1.1 Analysis of dashboard 1 (Midterm)
According to the dashboard of bar chart, representing the average of midterm by Student ID, it
was speculated that the highest score was S168, who scored 91 out of 100 in the midterm test
whereas the lowest score was S257, who scored 18 out of 100 in the midterm test. Hence, the
data also revealed that S168 is a computer science undergraduate student while S257 is a
business information system undergraduate student. To add on, only 5 out of 239 students
managed to achieve excellent mark (marks >= 75) which once again from a computer science
background which covers only 0.02% while the other 234 students (99.8%) scored below 75.
Although students from information systems engineering background didn’t manage to surpass
most of the computer science students, their results are doing quite well at satisfactory level
where the highest IA student(S236) scored is 72 marks which is 19 marks below the highest
CS student who got 91. On the other hand, IB student(S250) who scored the highest (53 marks)
in his programme was apparently the lowest comparing to the other students from IA and CS.

Based on the pie chart according to the average of midterm by programme, it was observed
that CS students scored the highest average which was 57.24 marks followed by IA (48.78)
and IB (30.57). This is also fair to justify that student from CS programme covers the highest
percentage based on the data which is 41.91% while IA and IB covers 35.71% and 22.38%
respectively. Hence, the observed table data shows that 170 students of all programmes have
passed the midterm by scoring 50 and above. This also means that 71.13% of all students
managed to pass the midterm where the highest recorded was 91 marks and the lowest was 50
marks. Although majority of the students may have passed, it was also found out that only 2
students from IB (S250 & S255) managed to pass by scoring 53 and 51 respectively. In this
matter, 2 out of 14 students (14%) from IB programme were scoring at satisfactory level while
majority of CS students managed to pass with students totalling 151 out of 189 which makes it
79% of the CS programme. Surprisingly, less than half of students from IA programme

56
Chapter 6

managed to score above 50 where only 17 out 36 students passed. This also means only 47%
students have satisfactory scores.

Based on the table data showing students scoring below 50, 69 students from all programmes
have failed the midterm test. This means that 28.87% of all students have failed the midterm.
Data shows that 38 students from CS, 19 students from IA, and 12 students from IB have
performed below satisfactory rate. IB students were performing at a bad state as majority of
the students have failed which covers 85% of the programme. Students from CS on the other
hand were doing fine as only 20% of the students in the programme have failed while IA
students were also not doing quite well as 52% have failed the midterm, which exceeds a little
more than half. The horizontal bar chart stating the number of genders who took the course
shows that 154 students who were males and 85 students who were females out of the 239
students. This is equivalent to 64% of males and 36% of females are taking the course where
there are more males than females by 28%. Next, the scatter chart stating average of midterm
shows that the average score by female is 54.64 and male is 54.28. This generally means that
the average of female scored higher than the average of male by 0.36 marks. Although there
are more male students who took this course than number of females, it was surprised that the
average of 85 females score higher than the average of 154 males.

Hence, according to the data from the line chart based on the average of midterm by gender
and programme, it shows that the average of female and male students from CS scored the
highest among all the programmes, 58.10 and 56.73 respectively. On the other hand, the
average of female and male students from IA scored averagely between CS and IB, 45 and
49.39 respectively. However, the average of female and male students from IB scored the
lowest among all programmes, averaging at 32.67 and 26.80 respectively where the students
scored below satisfactory level regardless of gender. Last, the tree map data shows that 189 CS
students, 36 IA students and 14 IB students out of 239 students are taking this midterm.
Generally speaking, CS, IA and IB are made up of 79%, 15%, and 5.8% respectively of the
whole programme.

57
Chapter 6

6.1.2 Analysis of dashboard 2 (Practical)

Based on the dashboard of bar chart which represents the average of practical by Student ID,
it was speculated that the highest score was S134, who scored 98 out of 100 in the practical
test whereas the lowest score was S244, who scored 2 out of 100 in the practical test. Hence,
the data also revealed that S134 is a computer science undergraduate student while S244 is an
information system engineering undergraduate student. To add on, 56 out of 239 students score
excellent marks which is 75 and above. This is a huge leap compared to the midterm where
only 5 students managed to score excellent marks. Also, all the 56 students are made from CS
students which is equivalent to 23% of the total of students all programmes. The highest marks
scored by an IB student(S251) was 72.50 which is slightly better than the previous result in
midterm while the highest marks score by an IA student was 73 marks which is also slightly
better than the previous midterm. Although IB student has the highest of 72.50, CS student
remains the all-time highest of 98 follows by IA which is 73.

According to the pie chart, stating the average of practical by programme, students from CS
programme scored the highest average, which is 62.91 followed by IB and IA, 44.29 and 41.69
respectively. This is equivalent to say that CS covers 42.25%, IB at 29.74% and IA at 28%. It
was analysed that the average of IA in practical test wasn’t performing as good as in midterm
which has the lowest average among all programmes. Next, the bar chart according to average
of practical test by gender shows that the average of male scored slightly higher than female
with scoring 59 for male while 58 for female. This is analysed that the average of 154 male
students score higher by 1 mark than the average of 85 females. Also, the horizontal bar chart
of max practical test by programme shows that CS student leads the highest score, which is 98
followed by IA, 73 while IB, 72.50. This justifies that CS students was far ahead than students
from the other programme. Hence, the funnel, according to the max practical test by gender
shows that the highest score by female and male student have the equal score which is 98. In
fact, both students were also students from CS programme.

Looking at the table data based on the number of students who passed the practical by scoring
50 and above, it was analysed that 158 students managed to pass. This means that 66% of the
students from all programmes are performing at a satisfactory result. Majority of the students
are once again from CS programme where it made out from 140 students while 12 students

58
Chapter 6

from IA while 6 students from IB. It was analysed that most of the IA and IB students didn’t
scored 50 and above as both measuring at 33% and 42% respectively which is not more than
half of the students in the same programme. Based on the table data showing students scoring
below 50, 81 students from all programmes have failed the midterm test which has higher
failure rate than in midterm. This means that 33% of all students have failed the midterm. Data
shows that 49 students from CS, 24 students from IA, and 8 students from IB have performed
below satisfactory rate. IA students were performing at a bad state as majority of the students
have failed which covers 66% of the programme. Students from CS on the other hand were
doing fine as only 25% of the students in the programme have failed while IB students were
also not doing quite well as 57% have failed the midterm, which exceeds a little more than half.
Lastly, based on the tree map data stating the median of practical by programme, CS remain
the highest median score of 63.50 followed by IA, 44 and IB, 38.50.

6.1.3 Analysis of dashboard 3 (Assignment)

Based on the bar chart of assignment by students, it was found out that S168 scored the highest
which is 84 while the S245 scored the lowest which is 63. S168 appeared to be from CS while
S245 is from IB. Looking on the pie chart data of average of assignment by programme, the
average of CS students obtained the highest score at 72.18 followed by IB students at 71.21
and IA students at 69.69. Next the stacked column chart shows the average of assignment
result by gender, male and female. It was analysed that the score of male is slightly higher than
female, in which the score of male at 71.81 while 71.65. This show the difference in 0.16
between both genders. Next, the table shows the data of students who passed the
assignment(=>50). It was discovered that none of the students failed the assignment with
passing rate of 100% regardless of programme or gender. Other than that, the clustered bar
chart shows the max of assignment scored by programme. The highest reported is 84 by CS
programme followed by IA at 80 and IB at 77. Meanwhile the minimum score of assignment
is 63 by IB programme followed by IA, which is also at 63 and followed by CS which is at 64.
The tree map shows the median of assignment by programme. The highest median scored was
at 72 by CS, followed by IB at 71 and lastly followed by IA at 70. Lastly, the max of assignment
shows that male scored slightly highly than female, in which male scored at 84 while female
scoring at 82.

59
Chapter 6

6.1.4 Analysis on dashboard 4 (Final score result)

Based on the clustered column chart on the finals score by StudID, the highest score obtained
is S191, scoring 90 while the lowest score obtained is S37, scoring 38. The four students who
scored 0 might have withdrawn from the course or might be absent for the final exam. Next,
the pie chart shows the average of finals score by programme. CS students have the highest
average score at 65.59, while IA scoring at 60.02 and IB scoring at 52.78. The table data which
shows students who passed the finals with a score(=>40). In total, 234 students have passed
the exam with a minimum score of 40 while 5 students have failed the exam, where one of the
students scored 38 while the others got 0. The clustered column chart shows the average of
final score by gender. It was examined that female students scored slightly higher than male
where female category scoring at 63.92 whereas male scoring at 63.01. The next tree map
shows the average of final exam Q1, Q2, Q3, Q4. The highest scored obtained is in final exam
Q1 at 19.71, followed by final exam Q3 at 15.82, final exam Q2 at 15.36 and final exam Q4 at
12.43. The clustered column chart of average of final exam Q1 by programme shows that the
highest average scored is by CS at 20.88 followed by IA at 19.72 and followed by IB at 3.99.
Next, the clustered column chart of average of final exam Q2 by programme shows that the
highest average scored is by CS at 16.35, followed by IA at 14.06 and IB at 5.29. The next
clustered column chart of average of final exam Q3 by programme shows that the highest
average scored is by IB at 21.36 followed by CS at 16.39 and IA at 10.72. The final clustered
column chart of average of final exam Q4 by programme shows the highest average scored by
IB at 29.29 followed by CS at 11.97 and AI at 8.28.

6.1.5 Analysis on dashboard 5 (Final result)

Based on the clustered column chart, it was analysed that the highest score obtained was by
student S168 at 88.20, from CS programme while the lowest obtained was student S211 at
20.85, from IA programme. The pie chart shows the average of final result by programme,
where the highest average score is 64.97 by CS programme followed by IB at 54.79 and IA at
54.25. Next, the table which shows the amount of students who passed this subjects shown that
223 students have passed while the other 16 students have failed the subject, scoring below 50.
This also means that 93% of the students have passed while the other 7% students have failed.
It was also found out that males have a higher failing rate compared to females. The clustered
column chart showing the average of final result by gender where female scored a higher

60
Chapter 6

average at 63.03 compared to male, scoring an average at 62.61. Next the clustered bar chart
shows that the maximum achieved was 88.20 by CS, followed by IB at 66.73 and IA at 66.05.
The tree map shows the median of final results based on programme achieved where the highest
is still obtained by CS students following IA and IB. Lastly, the clustered column chart shows
the comparison of final exam and coursework. It was analysed that students tend to score higher
at final exam, scoring at an average of 31.67 while the coursework mark at 31.09.

6.2 Discussion of results

Based on the findings after the performed analysis on the dashboard, it was found out that
students from CS background generally performed better than students from other programmes
followed by IA and IB. One of the very main reasons is students who studied computer science
are smarter as they do enjoy math and problem-solving skills where these are the requirements
for a subject that needs coding. In fact, computer science programme is not an easy path as it
requires much continuous coding practices every day. Also, many people believe computer
science to be amongst the most difficult subjects to study. They feel that Computer Science is
a challenging subject to study and there is no denying it. Caused by a variety of circumstances,
studying, and understanding computer science may be difficult. Numerous factors like some of
the most challenging areas in the discipline include computer technology, software, and
statistical techniques. So, meaning to say that website development wasn’t difficult to most CS
students as they had experienced in more harder subjects which require complicated coding
structures like C++ language. In this matter, website development is considered easier for them
who has deep experience in other programming languages. Besides, it was also found out that
students from CS have exposure to coding since foundation in science, before entering to
degree field. On the other hand, IA students did not perform as good as CS students since IA
has a narrower scope while CS has a wider scope. For instance, IA requires less details of
mathematical theories and less complexity compared to CS although it still requires maths and
problem-solving skills. Unfortunately, students from IB didn’t perform as its peak comparing
to CS and IA. One of the main factors is that the students never expose much in the field of
coding as it was also a business-related programme. In fact, students have less experience in
coding where what is learnt throughout the programme was broad and only basics of
programming was learnt, and with a mixture of business knowledge as well. This is also fair to
say that programming language which IB students are learning is the tip of the iceberg. Besides,

61
Chapter 6

IB students have no coding knowledge or experience as most of them enrolled into foundation
in arts leading to accountancy and management where no subjects required coding. In this
matter, it is studied that the students basically have zero exposure and very novel to them. It
was also studied that students from all programmes generally performed better in practical test
compared to midterm where the average score in midterm was 54.41 while average of practical
score is 58.63, which is 4.22 marks higher. It is also safer to say that most students managed to
pass both tests. Based on the bloom taxonomy, it was studied that students performed better in
psychomotor domain than in cognitive domain. This means that students are better at hands on
coding physically at computer lab rather than memorising or knowledge based. This is because
most students have hands on practice their coding during practical class session every week
and more time are spent on coding rather than reading. As midterm test will require knowledge
and understanding on the theoretical information, students didn’t manage to score higher as it
require much studying and reading to recall the information absorbed. Other than that, it was
also observed that the average female students scored higher than average of male in midterm
while average male students scored higher than average of female in practical. This also means
that females performed better in cognitive domain whereas males performed better in
psychomotor. One of the factors is that females are basically good at reading and writing while
males are better in terms of logical and problem solving. Cognitive tasks like sitting for a
midterm test, where memory derived from, grows, and develop more rapidly in females that in
males. As a result, this brings an impact to reading as well as writing skills. In terms of
psychomotor, where it requires logical thinking, this is where males come to place as males are
studied to have better logical reasoning when it comes to problem solving in practical. Hence,
it was analysed that there are more failures in practical test than in midterm test. This also
means that most of the students are performed better in cognitive than in psychomotor domain
although the average of practical is slightly higher than midterm. This explains that majority
of the students are reasonably good in knowledge-based task as most of the exams encountered
by students throughout the stages of education are mostly cognitive which includes quiz,
midterm, and finals and assignment. These covers 80% of the total course where cognitive
domain is involved. Another factor which leads to more failure rate in practical is that students
have shallow understanding of the programming concept through the act of personal experience
like practicing coding everyday instead of theoretical learning. This also justified that student
didn’t spend much time on coding practices which able to improve engagement as well as
knowledge retention and this is necessary in every subject that need coding skills. On the other
hand, it is analysed that there are 100% passing rate for the assignment task. This also means

62
Chapter 6

that most of the students have better cognitive skills by encouraging discussions of what is
being taught in class and to explore and understand of how the principles are to be applied into
assignment tasks. Besides, the final exam score(50%) also shows that there 98% passing rate
where only 5 students failed the test, out of 239 students. As for the final result, it was analysed
that 223 students have passed the course while the other 16 students have failed the course.
This also means that majority of the students have strong grasph of knowledge and skills about
this subject. Lastly, it was also found out that the final exam overall score of students were
slightly higher than their coursework marks. This is proven that students are better in terms of
cognitive domain than psychomotor domain.

6.3 Recommendations and formulations

First and foremost, one of the recommended solutions is that higher educational institution
should implement another foundation programme, leading to Information and Communication
Technology so that students are able to expose in the field of technological world before
enrolling into an IT course. It is vital to equip students with the basics of programming skills
which will be required and relevant during degree studies so that students won’t have troubles
or difficulties when learning the concepts of programming. It is because most students who
performed badly are majority from IB, where they have no certain knowledge of discipline in
its degree. By introducing a new IT programme in foundation, the subjects will tailor towards
enabling students to prepare for their undergraduate computing studies as students have to get
used to principles and concepts of the Information Technology industry. On the other hand, the
higher educational institution could also introduce a few subjects related to programming
language like Introduction to C++ or Java in the foundation of arts (Accountancy and
Management) programme to allow students to put an emphasis on basics of computing skills,
especially IB or IA students before enrolling into degree of IT. Furthermore, it is also
recommended that lecturers or tutors pay more attention to the students who lacks capabilities
in learning and performed below satisfactory level regardless of degree programme or gender
by asking the students in meeting to solve the questions of a particular problem instead of
implementing spoon feeding culture as it tends to create ‘learned helplessness’ among students.
This condition will results students into becoming reliant on others instead of solving by
themselves. When students can’t solve it, most of them gave up and apparently lose interest in

63
Chapter 6

the subject. Learning how to program will be difficult when there is no interest in it. And
because of this, helpless students will take very little responsibility in their own learning and
eventually lose interest in the subject and ended up failing the exam. In fact, listening in class
and reading concepts of coding is not enough to adopt the skills as coding requires mass number
of practices, effort, time, and self-experimenting. Besides, a development of programming
courseware is also a suggested solution to improve students’ weak performance. As most
students begin as a novel starter, they are encouraged to learn HTML and CSS before moving
on to harder programming languages like JavaScript as it don't need to have any kind of
programming experience and less complexity. Common languages like JavaScript might pose
difficulties to certain students due to its multi-paradigm nature and more advanced syntax. For
instance, the courseware will be offered to every student throughout the semester regardless of
weak or smart students. Also, this programming courseware will have different sets of lessons
covering different chapters, exercises, examples, and quizzes. In every chapter, each student
must complete the exercises and score within a set target by the lecturers to prove the
understanding on the chapters. This will allow lecturers to identify which area of weaknesses
are the student into. Another recommendation is by establishing a coding club in the faculty of
information and technology where it allows students from all IT programmes to develop
applications together as a team and discover latest software. Lecturers could step in as the
project managers by assisting students to guide them on the right track as well as diagnosing
bugs and fixes. This club would also enable students to think out of the box and exchanging
ideas on how to code an application as programming is more than coding itself. Lastly, lecturers
are also recommended to implement live coding teaching instead of using slides to explain the
concepts and principles. During live coding, lecturers should construct programs in front of
students and its much more effective for many factors. First, students get to watch how the
lecturer solves the problem and correct the mistakes. By doing this, students know what is right
to do and what is wrong to do so. This will help students into preventing mistakes when coding.
For example, lecturer demonstrates the execution of the program by adopting “what-if” to test
out what would be the output if different methods are used. At this point, students can learn
quicker, knowing the rules of thumb to avoid unanticipated bugs and errors. Live coding will
encourage students to make predictions the outcomes of the output for all programs run by the
lecturer. As a result, more students able to understand what is conveyed instead of using
traditional teaching method where lecturer will continuously be teaching without noticing
which students are left behind to complete the course syllabus. In a matter of fact, students who
can’t catch up with the syllabus will performed badly.

64
Chapter 7

Chapter 7
Conclusion and Recommendation

7.1 Project review and discussion

As a conclusion, the findings based on performed data analysis on the results of students have
been identified as well as the area of weaknesses of students. It is proven that business
intelligence plays a very significant role in education especially into improving students’
performance. By undergoing data analysis using BI tools, higher educational institution able to
use the collected data gathered and visualized them into a graphical representation to extract
valuable insights into solving the problem. And by having this information extracted, it enables
universities to perform better at important decision makings. The study findings given in this
research show that using Business Intelligence tools to analyse existing data and derive
valuable information for decision making has a lot to promise. Students are also expected to
improve their ways of learning instead of relying on the notes and learning materials provided
by lecturer. The materials itself are insufficient to gain full understanding on the subject. More
practice and self-experimenting on coding is required to equip with the knowledge to place
themselves more competitive. Furthermore, the bloom taxonomy contributed much beneficial
outcomes by outlining the learning objectives for students and lectures to further understand
the purpose of the course. With it, students’ weaknesses are justified based on the domains and
able to find out which area they poorly performed at. Besides, by getting additional data to the
original dataset, the student performance analysis might be significantly enhanced. For
instance, it will be informative to see and compare the results of how other students
performed in the other university courses they take during the semester to enhance the further
analysis and discovering greater insights. This will aid in a greater understanding of the
students' performance and learning capabilities. Finally, leading educational institutions
throughout the world must employ innovative methods to tackle the problems and new
possibilities that have arisen. Also, the overall research of this project has proven that gender,
programme studied, background experience have an significant impact on the students’
performance. The implementation of advanced analytical techniques, such as business
intelligence (BI) systems and data mining tools and techniques, allows for more substantial use
and analysis of obtainable university data, resulting in much more effective and efficient

65
Chapter 7

performance, efficient governance, and informed choices focusing on clean, precise, stable,
and consistent data.

7.2 Novelties and Contributions

First of all, one of the novelties of project is the implementation of python along with the use
of anaconda navigator and Jupyter Notebook as my platform to perform my data visualization.
This is to make use of programming tool to solve the problem rather than using in-built business
intelligence tools, which other projects have already been done by others before. One of the
very reasons why Python is the preferred choice for data analysis and visualization is due to its
basic and simple syntax, tons of frameworks, and cutting the time needed to write a line of
code, in which makes it an excellent programming tool. On the other hand, it provides support
for data science libraries like Keras, Matplotlib and TenserFlow as they are ideally made for
data gathering, analysis, visualization and data modelling. In other words, Python is an
extremely popular programming tool which is a current trending language for data science and
widely used globally. Besides, Jupyter Notebook was used as well as the platform for Python
due its great environment where to develop code and also interacting with the results. It doesn’t
only work as an integrated development environment(IDE) but also a form of presentation as
well.

Secondly, bloom taxonomy is used to in this project as an indicator to find out which area of
weaknesses are the students into. The Bloom taxonomy is classified into the different domains,
like cognitive, affective and psychomotor. For instance, cognitive domain is mainly focused
on skills like intellectual abilities including analytical analysis, conflict resolution, and
knowledge building. The affective domain emphasizes students' emotions, beliefs, passions,
and admiration. Its linked hierarchy starts with taking in there and paying attention to
information and progresses to description, or integrating values as well as continuously
upholding them. It aims to assist students in comprehending their personal values and how
they've evolved. While the includes all aspects of a student's physical capabilities for task
completion, mobility, plus skill performance. These would be a great way to evaluate student’s
performance in their academics, which what most of the projects lacked off. Data visualization
itself is not enough to achieve the results as it may not be accurately find out which area of
weaknesses are the students into.

66
Chapter 7

Lastly, one of the novelties is comparison of available business intelligence tools on the
markets. It might be difficult to compare BI tools. Particularly considering the abundance of
products currently available that all claim to offer a wide range of business advantages while
also providing solutions to a wide range of issues. With all the available tools on the market,
we must explore what is the current trending tool being adopted by most companies or
organization for decision making, business intelligence most important criteria’s, and to study
what are the latest features and the significance of them to the project. Therefore, it is vital to
do a comparison to find which tools may suit to solve the current problem.

The contributions of this project will assist lecturers in higher institution gain a better insight
and understanding on the academic performance of students as well as knowing which area of
weaknesses are they into so that students in higher education could be able to have a huge leap
in academic performance.

7.3 Future work

In the future work, more complex business intelligence tools will be used like Tableau as it is
a growing and trending tool implemented in the organizations for decisional making. Also,
there will be more novel technologies and features introduced like data governance, self-service
BI, collaborative and integrative BI and so many more. Therefore, to remain competitive in the
data and analytics effort, these trends must be adopted and embraced and strengthening today’s
competitive business world. Aside from that, existing sample data from other higher
institutions will be analysed as more meaningful data provides a better and concrete decision
making. This will includes data analysis on different subjects as well to find out more justifiable
information on the area of weaknesses the students are into. Also, Microsoft SQL Server would
be considered to be implemented in the future as the database because it is well known and
trending among data science industry and also extremely compatible and suited with Microsoft
Azure as well as Microsoft’s Business Intelligence tools. Hence, this database is designed to
help data analysts who need to access enormous datasets quickly and effectively and as large
amount of data will be needed in future project, this would be the right tool to use. SQL Server
has the capability to handle a variety of datasets, particularly non-relational but also
unstructured data, whereas most systems concentrate on managing organized and hierarchical
information.

67
REFERENCES

References
[1] R. Agarwal, V. Dhar, “big data, data science, and analytics: The opportunity and
challenge for IS research, Information Systems Research”, 2014, pp. 443-448.

[2] A. Shollo, K. Kautz, “Towards an understanding of business intelligence, In the

Proceedings of the 21st Australian Conference in Information Systems (ACIS)”, 2010.

[3] C.Carlos, M.Steven, R.Peter, “Database Systems: Design, Implementation, and

Management, Course Technology”, 2014, ch.13

[4] “The top 7 Benefits of Business Intelligence (BI)”. [Online] Available:

https://www.tableau.com/learn/articles/business-intelligence/enterprise-business-
intelligence/benefits Accessed: 11 March 2022

[5] Arpaci, I, “Understanding and predicting students’ intention to use mobile cloud
storage services”, 2016, pp.150–157.
[6] D. Apraxine and E. Stylianou, “Business Intelligence in a Higher Educational,”
2017, pp. 1735–1746.

[7] “What is Cloud BI,” Accessed on Mar. 11, 2022. [Online]. Available:
https://www.domo.com/glossary/what-is-cloud Accessed: 23 November 2021

[8] Debattista, J. Londoño, S. Lange, C. Auer, “A Framework for Linked Data Quality
Assessment”, 2016. [Online] Available:
https://www.researchgate.net/publication/269418019_Luzzu_- Accessed: 23
November 2021
[9] Gartner. (Apparently of 2018b), “IT Glossary”, 2016. [Online] Available:
https://www.gartner.com/it-glossary/self-service-business-intelligence Accessed: 24
November 2021
[10] Informatec, “Work more effectively: With Self-service Business Intelligence”,
2018. [Online] Available: https://www.informatec.com/en/effective-selfservice-bi
Accessed: 24 November 2021
[11] Norman, Patricia M, "Are Your Secrets Safe? Knowledge Protection in Strategic
Alliances." ,2001. [Online] Available:
https://www.sciencedirect.com/science/article/pii/S0007681301800732 Accessed: 14
November 2021
[12] Marelli, L. Testa, G, “Scrutinizing the EU General Data Protection Regulation”,
2018. [Online] Available:
https://www.researchgate.net/publication/324936977_Scrutinizing_the_EU_General_
Data_Protection_Regulation Accessed: 24 February 2022

68
REFERENCES

[13] Gartner. (January 2017), “Cloud Computing Enters its Second Decade”, 2017.
[Online] Available: https://www.gartner.com/smarterwithgartner/cloudcomputing-
enters-its-second-decade/ Accessed: 24 November 2021
[14] García, J./Technology Evaluation Centers, “Mobile BI: Features, Challenges, and
Opportunities”, 2010. [Online] Available:
https://www.whitepapers.em360tech.com/wp Accessed: 24 November 2021
[15] BARC, “The Business Application Research Center. (November 2018a). BI
Trend Monitor 2019”. [Online] Available: http://barcresearch.com/research/bi-trend-
monitor Accessed: 24 November 2021
[16] Bitner, S./Sisense, “Does the Future Lie with Embedded BI”. [Online] Available:
https://www.sisense.com/blog/future-lie-embedded-bi/ Accessed: 24 November 2021
[17] Dykes, B./Forbes, “Data Storytelling: The Essential Data Science Skill Everyone
Needs”, 2016 [Online] Available:
https://www.forbes.com/sites/brentdykes/2016/03/31/data-storytelling-theessential-
data-science-skill-everyone-needs/#42e8023752ad Accessed: 24 November 2021
[18] Micek, B./CARTO, “How Entrepreneurs are Using Open Data to Start
Businesses” , 2017 [Online] Available: https://carto.com/blog/howentrepreneurs-
using-open-data-start-businesses/ Accessed: 21 November 2021
[19] BARC, “The Business Application Research Center. (Apparently of 2018c). Data
Quality and Master Data Management: How to Improve Your Data Quality”, 2016
[Online] Available: https://bi survey.com/data-quality-master-data-management
Accessed: 21 November 2021
[20] Galetto, M./NGDATA, “What is Big Data Analytics?”, 2016, [Online]
Available: https://www.ngdata.com/what-is-big-data-analytics/ Accessed: 23
November 2021
[21] Rajesh, & Ramesh, “An introduction to data lake. I-Manager's Journal on
Information Technology, 5(2), 1”, 2016 [Online] Available:
http://www.splunk.com/web_assets/pdfs/secure/Splunk_for_BigData.pdf Accessed:
21 November 2021
[22] Shepherd, A. Kesa, C. Cooper, J. Onema, J. Kovacs, P, “Opportunities and
challenges associated with implementing Data lakes for enterprise decisionmaking”,
2018 [Online] Available: http://www.iacis.org/iis/2018/1_iis_2018_48-57.pdf
Accessed: 21 November 2021
[23] Cambria, Erik, “Affective computing and sentiment analysis. IEEE Intelligent
Systems”, 2016, sec 31, pp.102-107
[24] Chandrayan, P, “A Guide To NLP : A Confluence Of AI And Linguistics", 2017
[Online] Available: https://codeburst.io/a-guide-to-nlp-a-confluence-of-aiand-
linguistics 2786c56c0749 Accessed: 21 November 2021

69
REFERENCES

[25] B. Wieder , M.-L. Ossimitz, “The Impact of Business Intelligence on the Quality
of Decision Making – A Mediation Model,” Procedia Comput, vol. 64, 2015, pp.
1163–1171
[26] D. Arnott, F. Lizama, and Y. Song, “Patterns of business intelligence systems use
in organizations,” Decis. Support Syst., vol. 97, 2017, pp. 58–68.
[27] D. Apraxine and E. Stylianou, “Business Intelligence in a Higher Educational”,
2017, pp. 1735–1746.
[28] A. B. Sangar and N. B. a. Iahad, “Critical Factors that Affect the Success of
Business Intelligence Systems (BIS) Implementation in an Organization” Int. J. Sci.
Technol. Res., vol. 2, no. 2, pp. 176–180.
[29] N. M. A. and M. A. M. Ahmad Fadhil Yusof, Suraya Miskon, Norasnita Ahmad,
Rose Alinda Alias, Haslina Hashim, Norris Syed Abdullah, D 2015, “Implementation
issues affecting the business intelligence adoption in public university,” ARPN J. Eng.
Appl. Sci., vol. 10,no. 23.
[30] P. Wanda and S. Stian, “The Secret of my Success: An exploratory study of
Business Intelligence management in the Norwegian Industry,” Procedia Comput.
Sci., vol. 64, 2015, pp. 240–247.
[31] V. Trieu, “Getting value from Business Intelligence systems : A review and
research agenda,” Decis. Support Syst., vol. 93, 2017, pp. 111–124
[32] A. I. Magaireah, H. Sulaiman, and N. Ali, “Theoretical framework of critical
success factors (CSFs) for Business Intelligence (BI) System,” in 2017 8th
International Conference on Information Technology (ICIT), 2017, pp. 455–463.
[33] A. H. Anjariny, A. M. Zeki, and A. I. Abubakar, “The Mediating Effect of
Teamwork toward Organizational Readiness for Business Intelligence (BI)
Implementation,”, 2015, pp. 176–181.
[34] J.-P. Dawson, L., Van Belle, “Critical success factors for business intelligence in
the South African financial services sector,” SA J. Inf. Manag., vol. 15, no. 1, 2013,
pp. 1–12,
[35] Kartika.M, Prismahardi.R, Tresna.F,Kristya.M, “Business Intelligence For
Educational Institution : A Literature Review”, 2020, pp.22-25 [Online] Available:
https://www.ijconsist.org/index.php/ijconsist/article/view/32 Accessed: 21 November
2021
[36] DataDrivenCulture., [Online] Available from https://www.datadrivenculture.dk/
.Accessed: 21 November 2021
[37] Harris, D./Software Advice. (n.d.), “4 Emerging Use Cases for IoT Data
Analytics”. [Online] Available: https://www.softwareadvice.com/resources/iot-
dataanalytics-use-cases/ Accessed: 23 November 2021

70
REFERENCES

[38] TDWI. (n.d.), “TDWI GLOSSARY: What is a data lake?”.[Online] Available:

https://tdwi.org/portals/what-is-a-data-lake-definition.aspx Accessed: 23 November
2021

71
APPENDIX

Sample data of students for the course UCCD9999

72
APPENDIX

73
APPENDIX

74
APPENDIX

Code (dashboard 1 - midterm)

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('midterm',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('studID', 'midterm', data=sample_data_sorted_desc)
plt.title('Average of Midterm by StudId')
plt.xlabel('Student Id', size=15)
plt.ylabel('Average of midterm', size=15)
plt.legend(labels = ['Average midterm'])
#display plot
plt.show()

#piechart
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['midterm'].mean()
groupby_mean1 = sample_data.groupby(['Programme']).mean()
print ('Mean midterm: ' + str(mean1))
Programme_data = sample_data["Programme"]
midterm_data = sample_data["midterm"]
plt.show()
print(groupby_mean1)
#Output the figure of piechart
# Import libraries
import numpy as np
import matplotlib.pyplot as plt
# Creating dataset

75
APPENDIX

programme = ['IA', 'IB', 'CS']

data = [48.77, 30.57, 57.24]
# Creating explode data
explode = (0.1, 0.0, 0.2)
# Creating color parameters
colors = ( "orange", "cyan", "brown")
# Wedge properties
wp = { 'linewidth' : 1, 'edgecolor' : "green" }
# Creating autocpt arguments
def func(pct, allvalues):
absolute = int(pct / 100.*np.sum(allvalues))
return "{:.1f}%\n({:d} g)".format(pct, absolute)
# Creating plot
fig, ax = plt.subplots(figsize =(10, 7))
wedges, texts, autotexts = ax.pie(data,
autopct = lambda pct: func(pct, data),
explode = explode,
labels = programme,
shadow = True,
colors = colors,
startangle = 90,
wedgeprops = wp,
textprops = dict(color ="magenta"))
# Adding legend
ax.legend(wedges, programme,
title ="Programmes",
loc ="center left",
bbox_to_anchor =(1, 0, 0.5, 1))

plt.setp(autotexts, size = 8, weight ="bold")

ax.set_title("The Average of midterm by programmes")

# show plot
plt.show()

76
APPENDIX

#studentaboveOrEqualTo50
import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols =
['studID','midterm', 'Programme'], low_memory = False)
dataframe.head()
print(
dataframe[
(dataframe['midterm'] >= 50)
]
)

#studentbelow50
import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols =
['studID','midterm', 'Programme'], low_memory = False)
dataframe.head()
print(
dataframe[
(dataframe['midterm'] < 50)
]

#Count of midterm by gender

import pandas as pd
data = pd.read_csv('Sample_data_for_a_sample_course.csv')
data.info()
data['Gender'].value_counts(ascending=False)
Gender = ['Male','Female']
Midterm = [154, 85]
plt.barh(Gender, Midterm, color=['blue', 'red'])
plt.title('Count of midterm by gender')
plt.ylabel('Gender')

77
APPENDIX

plt.xlabel('Count of midterm')
plt.show()

#Scatter chart
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['midterm'].mean()
groupby_mean1 = sample_data.groupby(['Gender']).mean()
print ('Mean midterm: ' + str(mean1))
Programme_data = sample_data["Gender"]
midterm_data = sample_data["midterm"]
plt.show()
print(groupby_mean1)
x = ['female', 'male']
y = [54.63, 54.27]
plt.title('Average of Midterm by Gender')
plt.xlabel('Gender', size=15)
plt.ylabel('Average of midterm', size=15)
plt.scatter(x, y, color=['red', 'blue'])
plt.show()

#Area chart
import matplotlib.pyplot as plt
# line 1 points
x1 = ['CS', 'IA', 'IB']
y1 = [56.73, 49.39, 26.80]
# plotting the line 1 points
plt.plot(x1, y1, label = "Male")
# line 2 points
x2 = ['CS', 'IA', 'IB']
y2 = [58.10,45,32.67]
# plotting the line 2 points
plt.plot(x2, y2, label = "Female")

78
APPENDIX

plt.xlabel('Programme')
# Set the y axis label of the current axis.
plt.ylabel('Average of midterm')
# Set a title of the current axes.
plt.title('Average of midterm by programme and gender')
# show a legend on the plot
plt.legend()
# Display a figure.
plt.show()

#Treemap
import pandas as pd
data = pd.read_csv('Sample_data_for_a_sample_course.csv')
data.info()
data['Programme'].value_counts(ascending=False)
pip install squarify
import matplotlib.pyplot as plt
import squarify
sizes=[189, 36, 14]
label=["CS", "IA", "IB"]
squarify.plot(sizes=sizes, label=label, alpha=0.6 )
plt.title('Average of Midterm by Programme')
plt.axis('off')
plt.show()

Code (dashboard 2 - practical)

#Bar chart
import pandas as pd
from matplotlib import pyplot as plt

79
APPENDIX

sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#clustered bar chart
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Practical',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('studID', 'Practical', data=sample_data_sorted_desc, color = ['maroon'])
plt.title('Average of Practical by StudId')
plt.xlabel('Student Id', size=15)
plt.ylabel('Average of practical', size=15)
plt.legend(labels = ['Average practical'])
#display plot
plt.show()

#Pie chart
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Practical'].mean()
groupby_mean1 = sample_data.groupby(['Programme']).mean()
print ('Mean practical: ' + str(mean1))
Programme_data = sample_data["Programme"]
practical_data = sample_data["Practical"]
plt.show()
print(groupby_mean1)
#Output the figure of piechart
# Import libraries
import numpy as np
import matplotlib.pyplot as plt
# Creating dataset
programme = ['IA', 'IB', 'CS']
data = [41.69, 44.28, 62.91]

# Creating explode data

80
APPENDIX

explode = (0.1, 0.0, 0.2)

# Creating color parameters
colors = ( "blue", "yellow", "brown")
# Wedge properties
wp = { 'linewidth' : 1, 'edgecolor' : "green" }
# Creating autocpt arguments
def func(pct, allvalues):
absolute = int(pct / 100.*np.sum(allvalues))
return "{:.1f}%\n({:d} g)".format(pct, absolute)
# Creating plot
fig, ax = plt.subplots(figsize =(10, 7))
wedges, texts, autotexts = ax.pie(data,
autopct = lambda pct: func(pct, data),
explode = explode,
labels = programme,
shadow = True,
colors = colors,
startangle = 90,
wedgeprops = wp,
textprops = dict(color ="magenta"))
# Adding legend
ax.legend(wedges, programme,
title ="Programmes",
loc ="center left",
bbox_to_anchor =(1, 0, 0.5, 1))

plt.setp(autotexts, size = 8, weight ="bold")

ax.set_title("The Average of practical by programmes")

# show plot
plt.show()

#Bar chart for practical by gender

81
APPENDIX

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Practical'].mean()
groupby_mean1 = sample_data.groupby(['Gender']).mean()
print ('Mean midterm: ' + str(mean1))
Programme_data = sample_data["Gender"]
midterm_data = sample_data["Practical"]
plt.show()
print(groupby_mean1)
x = ['female', 'male']
y = [58, 59]
plt.title('Average of Practical by Gender')
plt.xlabel('Gender', size=15)
plt.ylabel('Average of practical', size=15)
plt.bar(x, y, color=['red', 'blue'])
plt.show()

#Horizontal bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
max1 = sample_data['Practical'].max()
groupby_max1 = sample_data.groupby(['Programme']).max()
print ('Max practical: ' + str(max1))
Programme_data = sample_data["Programme"]
practical_data = sample_data["Practical"]
plt.show()
print(groupby_max1)
x = ['CS', 'IA', 'IB']
y = [98, 73, 72.5]
plt.title('Max of Practical by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Max of Practical Test', size=15)

82
APPENDIX

plt.barh(x, y, color=['orange', 'blue', 'yellow'])

plt.show()

#Funnel
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
max1 = sample_data['Practical'].max()
groupby_max1 = sample_data.groupby(['Gender']).max()
print ('Max practical: ' + str(max1))
Gender_data = sample_data["Gender"]
practical_data = sample_data["Practical"]
plt.show()
print(groupby_max1)

!pip install plotly

import plotly.express as px
data = dict(
practical=[98, 98],
gender=["Male", "Female"])
fig = px.funnel(data, x='practical', y='gender', color = ['blue', 'red'])
fig.update_layout(
title={
'text' : "Max of Practical Test by Gender",
'y': 0.9,
'x': 0.5,
'xanchor': 'center',
'yanchor': 'top'})
fig.show()

#who pass practical over or equal 50 marks

import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols =
['Gender','Practical', 'Programme'], low_memory = False)

83
APPENDIX

dataframe.head()
print(
dataframe[
(dataframe['Practical'] >= 50)
]
)

#who fail practical below 50

import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols =
['Gender','Practical', 'Programme'], low_memory = False)
dataframe.head()
print(
dataframe[
(dataframe['Practical'] < 50)
]
)

#Treemap
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
median1 = sample_data['Practical'].median()
groupby_median1 = sample_data.groupby(['Programme']).median()
print ('Median practical: ' + str(median1))
Programme_data = sample_data["Programme"]
practical_data = sample_data["Practical"]
plt.show()
print(groupby_median1)
import matplotlib.pyplot as plt
import squarify
sizes=[63.5, 44, 38.5]
label=["CS", "IA", "IB"]
squarify.plot(sizes=sizes, label=label, alpha=0.6 )

84
APPENDIX

plt.title('Median of Practical by Programme')

plt.axis('off')
plt.show()

Code (dashboard 3 - Assignment)

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Assignment',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('studID', 'Assignment', data=sample_data_sorted_desc)
plt.title('Average of Assignment by StudId')
plt.xlabel('Student Id', size=15)
plt.ylabel('Average of Assignment', size=15)
plt.legend(labels = ['Average Assignment'])
#display plot
plt.show()

#PieChart
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Assignment'].mean()
groupby_mean1 = sample_data.groupby(['Programme']).mean()
print ('Mean assignment: ' + str(mean1))
Programme_data = sample_data["Programme"]
assignment_data = sample_data["Assignment"]
plt.show()

85
APPENDIX

print(groupby_mean1)
#Output the figure of piechart
# Import libraries
import numpy as np
import matplotlib.pyplot as plt
# Creating dataset
programme = ['IA', 'IB', 'CS']
data = [69.69, 71.21, 72.18]
# Creating explode data
explode = (0.1, 0.0, 0.2)
# Creating color parameters
colors = ( "yellow", "green", "brown")
# Wedge properties
wp = { 'linewidth' : 1, 'edgecolor' : "green" }
# Creating autocpt arguments
def func(pct, allvalues):
absolute = int(pct / 100.*np.sum(allvalues))
return "{:.1f}%\n({:d} g)".format(pct, absolute)
# Creating plot
fig, ax = plt.subplots(figsize =(10, 7))
wedges, texts, autotexts = ax.pie(data,
autopct = lambda pct: func(pct, data),
explode = explode,
labels = programme,
shadow = True,
colors = colors,
startangle = 90,
wedgeprops = wp,
textprops = dict(color ="magenta"))
# Adding legend
ax.legend(wedges, programme,
title ="Programmes",
loc ="center left",
bbox_to_anchor =(1, 0, 0.5, 1))

86
APPENDIX

plt.setp(autotexts, size = 8, weight ="bold")

ax.set_title("The Average of assignment by programmes")

# show plot
plt.show()

#vertical bar graph

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Assignment'].mean()
groupby_mean1 = sample_data.groupby(['Gender']).mean()
print ('Mean assignment: ' + str(mean1))
Programme_data = sample_data["Gender"]
assignment_data = sample_data["Assignment"]
plt.show()
print(groupby_mean1)
x = ['female', 'male']
y = [71.65, 71.81]
plt.title('Average of Assignment by Gender')
plt.xlabel('Gender', size=15)
plt.ylabel('Average of assignment', size=15)
plt.bar(x, y, color=['red', 'blue'])
plt.show()

#who pass practical over or equal 50 marks

import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols =
['Gender','Assignment', 'Programme'], low_memory = False)
dataframe.head()
print(

87
APPENDIX

dataframe[
(dataframe['Assignment'] >= 50)
]
)

#Horizontal bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
max1 = sample_data['Assignment'].max()
groupby_max1 = sample_data.groupby(['Programme']).max()
print ('Max assignment: ' + str(max1))
Programme_data = sample_data["Programme"]
assignment_data = sample_data["Assignment"]
plt.show()
print(groupby_max1)
x = ['CS', 'IA', 'IB']
y = [84, 80, 77]
plt.title('Max of Assignment by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Max of Assignment Test', size=15)
plt.barh(x, y, color=['orange', 'blue', 'yellow'])
plt.show()

#Horizontal bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
min1 = sample_data['Assignment'].min()
groupby_min1 = sample_data.groupby(['Programme']).min()
print ('Min assignment: ' + str(min1))
Programme_data = sample_data["Programme"]

88
APPENDIX

assignment_data = sample_data["Assignment"]
plt.show()
print(groupby_min1)
x = ['CS', 'IA', 'IB']
y = [64, 63, 63]
plt.title('Min of Assignment by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Min of Assignment Test', size=15)
plt.barh(x, y, color=['orange', 'blue', 'yellow'])
plt.show()

#Treemap
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
median1 = sample_data['Assignment'].median()
groupby_median1 = sample_data.groupby(['Programme']).median()
print ('Median practical: ' + str(median1))
Programme_data = sample_data["Programme"]
assignment_data = sample_data["Assignment"]
plt.show()
print(groupby_median1)
import matplotlib.pyplot as plt
import squarify
sizes=[72, 70, 71]
label=["CS", "IA", "IB"]
squarify.plot(sizes=sizes, label=label, alpha=0.6 )
plt.title('Median of Assignment by Programme')
plt.axis('off')
plt.show()

#Funnel

89
APPENDIX

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
max1 = sample_data['Assignment'].max()
groupby_max1 = sample_data.groupby(['Gender']).max()
print ('Max assignment: ' + str(max1))
Gender_data = sample_data["Gender"]
assignment_data = sample_data["Assignment"]
plt.show()
print(groupby_max1)

!pip install plotly

import plotly.express as px
data = dict(
Assignment=[84, 82],
gender=["Male", "Female"])
fig = px.funnel(data, x='Assignment', y='gender', color = ['male', 'female'])
fig.update_layout(
title={
'text' : "Max of Assignment by Gender",
'y': 0.9,
'x': 0.5,
'xanchor': 'center',
'yanchor': 'top'})
fig.show()

Code (dashboard 4 - Final score)

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order

90
APPENDIX

sample_data_sorted_desc= sample_data.sort_values('Final score',ascending=False)

#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('studID', 'Final score', data=sample_data_sorted_desc)
plt.title('Average of Final score by StudId')
plt.xlabel('Student Id', size=15)
plt.ylabel('Average of Final score', size=15)
plt.legend(labels = ['Average Final Score'])
#display plot
plt.show()

#PieChart
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Final score'].mean()
groupby_mean1 = sample_data.groupby(['Programme']).mean()
print ('Mean final score: ' + str(mean1))
Programme_data = sample_data["Programme"]
finalScore_data = sample_data["Final score"]
plt.show()
print(groupby_mean1)
#Output the figure of piechart
# Import libraries
import numpy as np
import matplotlib.pyplot as plt
# Creating dataset
programme = ['IA', 'IB', 'CS']
data = [52.78, 60.02, 65.59]
# Creating explode data
explode = (0.1, 0.0, 0.2)
# Creating color parameters
colors = ( "grey", "red", "brown")
# Wedge properties

91
APPENDIX

wp = { 'linewidth' : 1, 'edgecolor' : "green" }

# Creating autocpt arguments
def func(pct, allvalues):
absolute = int(pct / 100.*np.sum(allvalues))
return "{:.1f}%\n({:d} g)".format(pct, absolute)
# Creating plot
fig, ax = plt.subplots(figsize =(10, 7))
wedges, texts, autotexts = ax.pie(data,
autopct = lambda pct: func(pct, data),
explode = explode,
labels = programme,
shadow = True,
colors = colors,
startangle = 90,
wedgeprops = wp,
textprops = dict(color ="magenta"))
# Adding legend
ax.legend(wedges, programme,
title ="Programmes",
loc ="center left",
bbox_to_anchor =(1, 0, 0.5, 1))

plt.setp(autotexts, size = 8, weight ="bold")

ax.set_title("The Average of final score by programmes")

# show plot
plt.show()

#who pass finals over or equal 40 marks

import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols = ['Gender','Final
score', 'Programme'], low_memory = False)
dataframe.head()

92
APPENDIX

print(
dataframe[
(dataframe['Final score'] >= 40)
]
)

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Final Exam Q4',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('Programme', 'Final Exam Q4', data=sample_data_sorted_desc)
plt.title('Average of Final Exam Q4 by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Average of Final Exam Q4', size=15)
plt.legend(labels = ['Average Final Exam Q4'])
#display plot
plt.show()

#Treemap
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Final Exam Q1'].mean()
mean2 = sample_data['Final Exam Q2'].mean()
mean3 = sample_data['Final Exam Q3'].mean()
mean4 = sample_data['Final Exam Q4'].mean()
print ('Average final exam of Q1,Q2,Q3,Q4: ' + str(mean1) + str(mean2) + str(mean3) +
str(mean4))

93
APPENDIX

final1_data = sample_data["Final Exam Q1"]

final2_data = sample_data["Final Exam Q2"]
final3_data = sample_data["Final Exam Q3"]
final4_data = sample_data["Final Exam Q4"]
plt.show()
import matplotlib.pyplot as plt
import squarify
sizes=[19.71, 15.36, 15.82, 12.43]
label=["Q1", "Q2", "Q3", "Q4"]
squarify.plot(sizes=sizes, label=label, alpha=0.6 )
plt.title('Average of Final exam Q1, Q2, Q3, Q4')
plt.axis('off')
plt.show()

#vertical bar graph

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Final score'].mean()
groupby_mean1 = sample_data.groupby(['Gender']).mean()
print ('Mean Final score: ' + str(mean1))
Programme_data = sample_data["Gender"]
finalScore_data = sample_data["Final score"]
plt.show()
print(groupby_mean1)
x = ['female', 'male']
y = [63.92, 63.01]
plt.title('Average of Final score by Gender')
plt.xlabel('Gender', size=15)
plt.ylabel('Average of final score', size=15)
plt.bar(x, y, color=['red', 'blue'])
plt.show()

94
APPENDIX

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Final Exam Q1',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('Programme', 'Final Exam Q1', data=sample_data_sorted_desc, color = ['green'])
plt.title('Average of Final Exam Q1 by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Average of Final Exam Q1', size=15)
plt.legend(labels = ['Average Final Exam Q1'])
#display plot
plt.show()

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Final Exam Q2',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('Programme', 'Final Exam Q2', data=sample_data_sorted_desc, color = ['yellow'])
plt.title('Average of Final Exam Q2 by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Average of Final Exam Q2', size=15)
plt.legend(labels = ['Average Final Exam Q2'])
#display plot
plt.show()

95
APPENDIX

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Final Exam Q3',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('Programme', 'Final Exam Q3', data=sample_data_sorted_desc, color = ['brown'])
plt.title('Average of Final Exam Q3 by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Average of Final Exam Q3', size=15)
plt.legend(labels = ['Average Final Exam Q3'])
#display plot
plt.show()

Code (dashboard 5 – final result)

#clustered bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
#sort data in descending order
sample_data_sorted_desc= sample_data.sort_values('Final result(100%)',ascending=False)
#specify width and height of figure
plt.figure(figsize = (13,7))
plt.bar('studID', 'Final result(100%)', data=sample_data_sorted_desc)
plt.title('Average of Final score by StudId')
plt.xlabel('Student Id', size=15)
plt.ylabel('Average of Final result', size=15)
plt.legend(labels = ['Average Final Result'])
#display plot
plt.show()

96
APPENDIX

#PieChart
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Final result(100%)'].mean()
groupby_mean1 = sample_data.groupby(['Programme']).mean()
print ('Mean final result: ' + str(mean1))
Programme_data = sample_data["Programme"]
finalScore_data = sample_data["Final result(100%)"]
plt.show()
print(groupby_mean1)
#Output the figure of piechart
# Import libraries
import numpy as np
import matplotlib.pyplot as plt
# Creating dataset
programme = ['IA', 'IB', 'CS']
data = [54.25, 54.79, 64.97]
# Creating explode data
explode = (0.1, 0.0, 0.2)
# Creating color parameters
colors = ( "grey", "red", "brown")
# Wedge properties
wp = { 'linewidth' : 1, 'edgecolor' : "green" }
# Creating autocpt arguments
def func(pct, allvalues):
absolute = int(pct / 100.*np.sum(allvalues))
return "{:.1f}%\n({:d} g)".format(pct, absolute)
# Creating plot
fig, ax = plt.subplots(figsize =(10, 7))
wedges, texts, autotexts = ax.pie(data,
autopct = lambda pct: func(pct, data),

97
APPENDIX

explode = explode,
labels = programme,
shadow = True,
colors = colors,
startangle = 90,
wedgeprops = wp,
textprops = dict(color ="magenta"))
# Adding legend
ax.legend(wedges, programme,
title ="Programmes",
loc ="center left",
bbox_to_anchor =(1, 0, 0.5, 1))

plt.setp(autotexts, size = 8, weight ="bold")

ax.set_title("The Average of final result by programmes")

# show plot
plt.show()

#who pass finals over or equal 50 marks

import pandas as pd
dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols = ['Gender','Final
result(100%)', 'Programme'], low_memory = False)
dataframe.head()
print(
dataframe[
(dataframe['Final result(100%)'] >= 50)
]
)

#who fail finals below 50 marks

import pandas as pd

98
APPENDIX

dataframe = pd.read_csv('Sample_data_for_a_sample_course.csv', usecols = ['Gender','Final

result(100%)', 'Programme'], low_memory = False)
dataframe.head()
print(
dataframe[
(dataframe['Final result(100%)'] < 50)
]
)

#vertical bar graph

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Final result(100%)'].mean()
groupby_mean1 = sample_data.groupby(['Gender']).mean()
print ('Mean Final result: ' + str(mean1))
Programme_data = sample_data["Gender"]
finalScore_data = sample_data["Final result(100%)"]
plt.show()
print(groupby_mean1)
x = ['female', 'male']
y = [63.03, 62.61]
plt.title('Average of Final result by Gender')
plt.xlabel('Gender', size=15)
plt.ylabel('Average of final result', size=15)
plt.bar(x, y, color=['red', 'blue'])
plt.show()

#Horizontal bar chart

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')

99
APPENDIX

max1 = sample_data['Final result(100%)'].min()

groupby_max1 = sample_data.groupby(['Programme']).max()
print ('Max final result: ' + str(max1))
Programme_data = sample_data["Programme"]
FinalResult_data = sample_data["Final result(100%)"]
plt.show()
print(groupby_max1)
x = ['CS', 'IA', 'IB']
y = [88.20, 63.73, 63.05]
plt.title('Max of final result by Programme')
plt.xlabel('Programme', size=15)
plt.ylabel('Max of final result', size=15)
plt.barh(x, y, color=['pink', 'purple', 'yellow'])
plt.show()

#Treemap
import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
median1 = sample_data['Final result(100%)'].median()
groupby_median1 = sample_data.groupby(['Programme']).median()
print ('Median final result: ' + str(median1))
Programme_data = sample_data["Programme"]
FinalResult_data = sample_data["Final result(100%)"]
plt.show()
print(groupby_median1)
import matplotlib.pyplot as plt
import squarify
sizes=[66.70, 56.38, 53.15]
label=["CS", "IA", "IB"]
squarify.plot(sizes=sizes, label=label, alpha=0.6 )
plt.title('Median of Final result by Programme')
plt.axis('off')

100
APPENDIX

plt.show()

#vertical bar graph

import pandas as pd
from matplotlib import pyplot as plt
sample_data = pd.read_csv('Sample_data_for_a_sample_course.csv')
mean1 = sample_data['Final result(100%)'].mean()
mean2 = sample_data['Coursework(50%)'].mean()
print ('Mean Final result: ' + str(mean1))
print ('Mean Coursework: ' + str(mean2))
finalScore_data = sample_data["Final result(100%)"]
coursework_data = sample_data["Coursework(50%)"]
plt.show()
x = ['coursework', 'final exam']
y = [31.09, 31.67]
plt.title('Average of Final result and coursework')
plt.ylabel('Average of final result and coursework', size=15)
plt.bar(x, y, color=['green', 'lightgreen'])
plt.show()

101
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 2

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Drawn the block diagram to convey the representation of the data analysis system
in the form of blocks

2. WORK TO BE DONE

- Review the literature review and cite articles

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project

4. SELF EVALUATION OF THE PROGRESS

- Just started to revise the previous project from FYP1 and continuing the
remaining tasks

_________________________ _________________________
Supervisor’s signature Student’s signature

102
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 3

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Do a literature review on the cite some articles related to business intelligence

2. WORK TO BE DONE

- Gather the sample raw data and visualize them into form of graphical
representations by integrating into Power BI software

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project

4. SELF EVALUATION OF THE PROGRESS

- Slow progress but at a constant rate

_________________________ _________________________
Supervisor’s signature Student’s signature

103
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 4

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Successfully visualized the data in the form of graphical representations

2. WORK TO BE DONE

- Visualize the data by integrating into Python to generate existing results

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project

4. SELF EVALUATION OF THE PROGRESS

- 50% completion of progress and the still in mid-way to finish the project

_________________________ _________________________
Supervisor’s signature Student’s signature

104
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 5

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Successfully output the results by using matplotlib and pandas libraries of Python

2. WORK TO BE DONE

- Debugging and testing the code

3. PROBLEMS ENCOUNTERED

- Some codes have minor errors and bugs

4. SELF EVALUATION OF THE PROGRESS

- Slow progress and the completion of the project is at 60%

_________________________ _________________________
Supervisor’s signature Student’s signature

105
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 6

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Debugged the codes and ensure all problems are fixed

2. WORK TO BE DONE

- Construct the dashboards of Power BI to provide critical reporting better

enhancement of analysis

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Slow but in progress and tasks are done daily

_________________________ _________________________
Supervisor’s signature Student’s signature

106
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 7

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Constructed the dashboard needed for Power BI showing the data of assignment
result, final exam and overall results.

2. WORK TO BE DONE

- Construct the dashboard for Python to provide critical reporting better

enhancement of analysis

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Slow but in still progress and tasks are done daily

_________________________ _________________________
Supervisor’s signature Student’s signature

107
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 8

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Constructed the dashboard for Python with all the required information needed

2. WORK TO BE DONE
- Analyze and discuss the results gathered based on the data shown in dashboard

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Project 80% completion and the visualization using Power BI is completed

_________________________ _________________________
Supervisor’s signature Student’s signature

108
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 9

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- The data is successfully analyzed with extracted information regarding the

performance of student results

2. WORK TO BE DONE

- Provide recommendations to solve the problems identified throughout the analysis

3. PROBLEMS ENCOUNTERED

- Some minor bugs encountered and some errors faced while trying to code out the
output.

4. SELF EVALUATION OF THE PROGRESS

- Project almost completed and there are only a few tasks left

_________________________ _________________________
Supervisor’s signature Student’s signature

109
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 10

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Recommendations are stated in the purpose to improve students’ performance

based on the weaknesses identified

2. WORK TO BE DONE

- Conclude the overall results throughout the project

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Project almost completed

_________________________ _________________________
Supervisor’s signature Student’s signature

110
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 11

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Conclusion has been completed based on the overview of the project

2. WORK TO BE DONE

- Novelties and contribution of the project

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Project almost completed

_________________________ _________________________
Supervisor’s signature Student’s signature

111
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 12

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Proposed the novelties of the project and what is its contribution

2. WORK TO BE DONE

- Design a poster to summarize the entire project. This will include the
introduction, objective, methodology, results, analysis, recommendations and
conclusion

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Project almost completed and only a few tasks left to go

_________________________ _________________________
Supervisor’s signature Student’s signature

112
APPENDIX

FINAL YEAR PROJECT WEEKLY REPORT

(Project II)

Trimester, Year: Trimester 3, Year 4 Study week no.: 13

Student Name & ID: Cheah Cheng Qian 17ACB04662
Supervisor: Ts Dr Chan Lee Kwun
Project Title: A Visualized Analysis on Student Performance by Using Cloud
Business Intelligence

1. WORK DONE
[Please write the details of the work done in the last fortnight.]

- Designed a poster for the project which summarizes the entire content of the
project report
- Format the report

2. WORK TO BE DONE

- The project is completed and thus, there are no tasks left to be done.

3. PROBLEMS ENCOUNTERED

- There are no problems encountered so far upon completing the project.

4. SELF EVALUATION OF THE PROGRESS

- Project has been completed

_________________________ _________________________
Supervisor’s signature Student’s signature

113
APPENDIX

POSTER

114
PLAGIARISM CHECK RESULT

PLAGIARISM CHECK RESULT

115
PLAGIARISM CHECK RESULT

Universiti Tunku Abdul Rahman

Form Title : Supervisor’s Comments on Originality Report Generated by Turnitin
for Submission of Final Year Project Report (for Undergraduate Programmes)
Form Number: FM-IAD-005 Rev No.: 0 Effective Date: 01/10/2013 Page No.: 1of 1

FACULTY OF INFORMATION AND COMMUNICATION

TECHNOLOGY

Full Name(s) of CHEAH CHENG QIAN

Candidate(s)
ID Number(s) 17ACB04662
Programme / Course BACHELOR OF INFORMATION SYSTEMS (HONOURS) BUSINESS
INFORMATION SYSTEMS
Title of Final Year Project A VISUALIZED ANALYSIS ON STUDENT PERFORMANCE BY USING
CLOUD BUSINESS INTELLIGENCE

Similarity Supervisor’s Comments

(Compulsory if parameters of originality exceeds
the limits approved by UTAR)

Overall similarity index: _10 %

Similarity by source
Internet Sources: ________9_____%
Publications: ___4___ %
Student Papers: ____0___ %

Number of individual sources listed of

more than 3% similarity: 0%
Parameters of originality required and limits approved by UTAR are as Follows:
(i) Overall similarity index is 20% and below, and
(ii) Matching of individual sources listed must be less than 3% each, and
(iii) Matching texts in continuous block must not exceed 8 words
Note: Parameters (i) – (ii) shall exclude quotes, bibliography and text matches which are less than 8 words.

Note Supervisor/Candidate(s) is/are required to provide softcopy of full set of the originality report
to Faculty/Institute

Based on the above results, I hereby declare that I am satisfied with the originality of the Final
Year Project Report submitted by my student(s) as named above.

____________________________ ______________________________
Signature of Supervisor Signature of Co-Supervisor

Name: Ts Dr Chan Lee Kwun Name: __________________________

Date: ____02 September 2022_____ Date: ___________________________

116
FYP2 CHECKLIST

UNIVERSITI TUNKU ABDUL RAHMAN

FACULTY OF INFORMATION & COMMUNICATION TECHNOLOGY
(KAMPAR CAMPUS)
CHECKLIST FOR FYP2 THESIS SUBMISSION
Student Id 17ACB04662
Student Name CHEAH CHENG QIAN
Supervisor Name TS DR CHAN LEE KWUN

TICK (√) DOCUMENT ITEMS

Your report must include all the items below. Put a tick on the left column after you have
checked your report with respect to the corresponding item.
Front Plastic Cover (for hardcopy)
√ Title Page
√ Signed Report Status Declaration Form
√ Signed FYP Thesis Submission Form
√ Signed form of the Declaration of Originality
√ Acknowledgement
√ Abstract
√ Table of Contents
√ List of Figures (if applicable)
√ List of Tables (if applicable)
List of Symbols (if applicable)
√ List of Abbreviations (if applicable)
√ Chapters / Content
√ Bibliography (or References)
√ All references in bibliography are cited in the thesis, especially in the chapter
of literature review
√ Appendices (if applicable)
√ Weekly Log
√ Poster
√ Signed Turnitin Report (Plagiarism Check Result - Form Number: FM-IAD-005)
√ I agree 5 marks will be deducted due to incorrect format, declare wrongly the
ticked of these items, and/or any dispute happening for these items in this
report.
*Include this form (checklist) in the thesis (Bind together as the last page)
I, the author, have checked and confirmed all the items listed in the table are included in my
report.

______________________
(Signature of Student)
Date: 02 September 2022

117
FYP2 CHECKLIST

118