[Grab your reader’s attention with a

great quote from the document or

use this space to emphasize a key
point. To place this text box
anywhere on the page, just drag it.]

ABSTRACT

Page 1 of 119
Art and music are essential aspects of human life. Humanity and art can't exist without each other. Sound
contact is unavoidable, whether you're making it or enjoying it. People have long valued music in their
lives, whether for listening pleasure, emotional response, performance, or creation. Music can stimulate
the brain as well. Music is extremely valuable at all stages of education. As music improves students'
minds, expressive abilities, and a variety of other attributes, they learn many vital and crucial life
values. The consumption of music on modern online streaming sites has become largely determined by
playlists, i.e., sequences of tracks that are intended to be listened to together (-Bonnin and
Jannach 2014; Schedl et al. 2018).

Humans listen to different types of music depending on their mood. Music has the ability to lift someone's
spirits, stimulate them, or soothe and rest them. It enables them to feel almost, if not all, of the emotions
that they encounter in their daily lives (-Galindo, G. 2003). However, people tend to get lazy when it
comes to finding new songs or discovering upcoming artists that they might possibly like.

For example, when someone is feeling low and wants to listen to an upbeat song to make themselves
feel better, they search for a song on different platforms, which proves to be very time-consuming, leaving
them to be even more frustrated. Additionally, if they are in a bad mood and put their music playlist on
shuffle, they might come across a song that might not match their mood, and they wouldn’t want to spend
a lot of time searching for the “perfect song” that fits the way they are currently feeling.

This report describes and analyzes the importance of recommended playlists, how automatically
generated playlists make life easier, and the importance of how mood affects a person’s music taste and
vice versa. Based on research and analysis, a website is being created to recommend users a playlist
based on their mood and their artist of choice whilst additionally suggesting different artists for the users
to listen to.

ACKNOWLEDGMENT

Before the start of this report, I would like to thank my project supervisor Ms. Geethu Joy for constantly
holding meetings on Teams throughout the different stages of the individual project and explaining what
the main deliverables are and giving advice. I would also like to thank Dr. Fehmida Hussain for holding
drop-in sessions and classes to explain how to go about with the project, report, and more.

Page 2 of 119
 TABLE OF CONTENTS
Serial Content Page Number
Number
1 Chapter 1 - Introduction 9
1.1 Introduction to the Project 9
1.2 Project Objective 11
1.3 Project Aim 11
1.4 Project Deliverables 11
1.5 Problems Addressed 12
2 Chapter 2 – Literature Review 12
2.1 Playlisting Introduction 12
2.2 Peculiarity of Music Recommendations 13
2.3 Automated Playlist Generation 13
2.3.1 Recommender Systems 13
2.3.2 Music Information Retrieval 15
2.4 Most Commonly Used Playlist Recommendation Techniques 16
2.4.1 Playlist Recommendation Based on Real-Time Behavior 16
2.4.2 Playlist Recommendation Based on Long-Term Behavior 17
2.5 Apple Music vs Spotify 17
2.6 Target characteristics for Music Playlist Generation 20
2.7 Difficulties Faced in Automated Playlist Generation 20
2.8 Research Methodology or Techniques 21
2.9 Research Methods 21
2.10 Research Findings and Analysis (Results) 22
3 Chapter 3 – Data Analysis 22
3.1 Introduction to the data being collected 22
3.2 Data Collection Process 22
3.3 Research Objectives 23
3.4 Initial Key Findings 23
3.5 Analysis on Mood with Genres 27
3.6 Analysis on Mood with Artists 29
3.7 Analysis on Artists 30
3.7.1 Pop Artist Recommendation 30
3.7.2 Hip-Hop/Rap Artist Recommendation 33
3.7.3 DJ Recommendation 34
3.7.4 Band Recommendation 36
3.8 Data Limitations 38
4 Chapter 4 – Requirements Specifications 38
4.1 Tools & Requirement Techniques 38
4.2 Functional Requirements 38
4.3 Non-Functional Requirements 39
4.4 User Interface Requirements 39
4.5 Software Requirements 39
5 Chapter 5 – Analysis and Design for the Website 40
5.1 Use Case Diagram 40
5.2 E/R Diagram 41
5.3 Activity Diagram 42
5.4 Features of the Website (Storyboard) 44
5.5 Wireframes of Initial Design Stages 45
5.6 Website Implementation and Source Code 50
5.7 Website Functionality & Final Design 57

Page 3 of 119
 5.7.1 Functionality of the Website 57
5.7.2 Website Final Design 57
6 Chapter 6 – Testing and Evaluation 78
6.1 Testing Checklist 78
6.2 Functionality Testing 79
6.3 Heuristic Evaluation 79
6.3.1 Procedure of Heuristic Testing 80
6.3.2 Results of Heuristic Testing 80
6.4 Usability Testing 84
6.4.1 Procedure of Usability Testing 84
6.4.2 Analysis on Behavior and Feedback 84
6.4.3 User Satisfaction Survey 85
7 Chapter 7 – User Guide 85
7.1 User Guide for Moodify 86
8 Chapter 8 – Conclusions 99
8.1 Evaluations & Summary 99
8.2 Limitations 99
8.3 Recommendations and Future Plans 100
9 Chapter 9 – References 101
A Chapter 10 - Appendix 106
A1 Supervisor Meeting Logs 107
A2 Source Code 106
A3 Selenium IDE Testing 106
A4 Ethical Screening Form 107
A5 Ethical Approval Form 109
A6 Research Phase Consent Form 111
A7 Usability Testing Consent Form 113
A8 Heuristics Evaluation Consent Form 114
A9 Moodify User Satisfaction Survey 115
A10 Moodify User Satisfaction Survey Response 116

TABLE OF FIGURES
FIGURE 1 - MUSIC CONSUMPTION STATISTICS ..........................................................................................................................8
FIGURE 2 - MOST IMPORTANT FEATURES OF STREAMING SERVICES ..............................................................................................9
FIGURE 3 - MUSIC RECOMMENDER SYSTEM ..........................................................................................................................12

Page 4 of 119
FIGURE 4 - MIR PROCESS ..................................................................................................................................................13
FIGURE 5 – SPOTIFY PERSONALIZED PLAYLISTS .......................................................................................................................16
FIGURE 6 - SPOTIFY PLAYLIST BASED ON USER HISTORY.............................................................................................................16
FIGURE 7 - TYPES OF PERSONALIZED PLAYLIST CURATED BY SPOTIFY ...........................................................................................17
FIGURE 8 - COMMON PLAYLISTS BY APPLE MUSIC ..................................................................................................................17
FIGURE 9 - APPLE MUSIC VS SPOTIFY ...................................................................................................................................18
FIGURE 10 - MUSIC LISTENING PATTERN...............................................................................................................................22
FIGURE 11 - MOOD LIFT UP BY MUSIC .................................................................................................................................22
FIGURE 12 - GENRE, MUSIC APP AND PREFERENCE.................................................................................................................23
FIGURE 13 - MOST COMMON MOOD ..................................................................................................................................23
FIGURE 14 - FAVORITE POP ARTISTS ....................................................................................................................................24
FIGURE 15 - FAVORITE RAP ARTISTS.....................................................................................................................................24
FIGURE 16 - FAVORITE BANDS ............................................................................................................................................25
FIGURE 17 - FAVORITE DJS.................................................................................................................................................25
FIGURE 18 - GENRE WHEN ANGRY ......................................................................................................................................26
FIGURE 19 - GENRE WHEN HAPPY.......................................................................................................................................26
FIGURE 20 - GENRE WHEN SAD ..........................................................................................................................................27
FIGURE 21 - MOST FREQUENTLY LISTENED TO ARTIST (HAPPY MOOD) .......................................................................................27
FIGURE 22 - MOST FREQUENTLY LISTENED TO ARTIST (SAD MOOD) ..........................................................................................28
FIGURE 23 - ALICIA KEYS....................................................................................................................................................29
FIGURE 24 - SHAWN MENDES & TAYLOR SWIFT.....................................................................................................................29
FIGURE 25 - KATY PERRY & ARIANA GRANDE ........................................................................................................................29
FIGURE 26 - DUA LIPA.......................................................................................................................................................30
FIGURE 27 - ED SHEERAN...................................................................................................................................................30
FIGURE 28 - CARDI B ........................................................................................................................................................31
FIGURE 29 – DRAKE ..........................................................................................................................................................31
FIGURE 30 - NICKI MINAJ ..................................................................................................................................................32
FIGURE 31 - ALAN WALKER ................................................................................................................................................32
FIGURE 32 – AVICII ...........................................................................................................................................................33
FIGURE 33 - DJ KHALED ....................................................................................................................................................33
FIGURE 34 - DJ SNAKE ......................................................................................................................................................33
FIGURE 35 - 5 SECONDS OF SUMMER ...................................................................................................................................34
FIGURE 36 – BTS .............................................................................................................................................................34
FIGURE 37 - ONE DIRECTION ..............................................................................................................................................35
FIGURE 38 - ONE REPUBLIC AND TWENTY-ONE PILOTS............................................................................................................35
FIGURE 39 - USE CASE DIAGRAM ........................................................................................................................................38
FIGURE 40 - ENTITY RELATIONSHIP DIAGRAM ........................................................................................................................39
FIGURE 41 - ACTIVITY DIAGRAM..........................................................................................................................................40
FIGURE 42 - MOODIFY STORYBOARD ...................................................................................................................................42
FIGURE 43 - INITIAL HOME PAGE DESIGN..............................................................................................................................43
FIGURE 44 - INITIAL REGISTER DESIGN WIREFRAME ................................................................................................................44
FIGURE 45 - INITIAL SIGN IN WIREFRAME ..............................................................................................................................44
FIGURE 46 - INITIAL ABOUT US DESIGN ................................................................................................................................45
FIGURE 47 - INITIAL GENRE SELECTION DESIGN ......................................................................................................................45
FIGURE 48 - INITIAL ARTIST OF CHOICE DESIGN ......................................................................................................................46
FIGURE 49 - INITIAL MOOD CHOICE DESIGN ..........................................................................................................................46
FIGURE 50 - CREATED USER PLAYLIST INITIAL DESIGN..............................................................................................................47
FIGURE 51 - MOODIFY.HTML ..............................................................................................................................................48
FIGURE 52 - OPENING SONG JAVASCRIPT FUNCTION ...............................................................................................................48

Page 5 of 119
FIGURE 53 - FUNCTION STORE ............................................................................................................................................49
FIGURE 54 - MOVING BACKGROUND....................................................................................................................................50
FIGURE 55 - AUTO PLAY SLIDESHOW ....................................................................................................................................50
FIGURE 56 - FUNCTION TO START THE PLAYLIST......................................................................................................................51
FIGURE 57 - HTML FOR PLAYLIST CREATION ..........................................................................................................................52
FIGURE 58 - ARTIST SUGGESTIONS .......................................................................................................................................53
FIGURE 59 - SAVING PLAYLIST CODE ....................................................................................................................................53
FIGURE 60 - MOODIFY HOME PAGE #1 ................................................................................................................................55
FIGURE 61 - TRAILER VIDEO ON HOMEPAGE ..........................................................................................................................55
FIGURE 62 - MOODIFY: NEW MUSIC THIS WEEK....................................................................................................................56
FIGURE 63 - MOODIFY: NEW ALBUMS RELEASED ...................................................................................................................56
FIGURE 64 - MOODIFY: COMING SOON ................................................................................................................................57
FIGURE 65 - MOODIFY FOOTER ...........................................................................................................................................57
FIGURE 66 - MOODIFY: REGISTER ........................................................................................................................................58
FIGURE 67 - MOODIFY: LOG IN ...........................................................................................................................................58
FIGURE 68 - MOODIFY: LEARN MORE ..................................................................................................................................59
FIGURE 69 - MOODIFY: LEARN MORE #2 .............................................................................................................................59
FIGURE 70 - DISCOVER PAGE WHEN OPENED.........................................................................................................................60
FIGURE 71 - DISCOVER PAGE WHILE SCROLLING .....................................................................................................................60
FIGURE 72 - DISCOVER POP ARTISTS ....................................................................................................................................61
FIGURE 73 - DISCOVER HIP HOP/RAP ARTISTS .......................................................................................................................61
FIGURE 74 - DETAILED ARTIST DISCOVER PAGE ......................................................................................................................62
FIGURE 75 - DISCOVER: ALBUMS RELEASED...........................................................................................................................62
FIGURE 76 - DISCOVER ALBUM BASED ON MOOD SUGGESTIONS ...............................................................................................63
FIGURE 77 - DISCOVER: SINGLES RELEASED ...........................................................................................................................63
FIGURE 78 - DISCOVER: LATEST PERFORMANCE .....................................................................................................................64
FIGURE 79 - DISCOVER: KATY PERRY ....................................................................................................................................64
FIGURE 80 - MAKE A PLAYLIST: IMAGE SLIDESHOW.................................................................................................................65
FIGURE 81 - MAKING A PLAYLIST .........................................................................................................................................65
FIGURE 82 - CREATE PLAYLIST: STEP 1..................................................................................................................................66
FIGURE 83 - CREATE PLAYLIST: SELECT GENRE .......................................................................................................................66
FIGURE 84 - CREATE PLAYLIST: SELECT POP ARTIST .................................................................................................................67
FIGURE 85 - CREATE PLAYLIST: SELECT BAND.........................................................................................................................67
FIGURE 86 - CREATE PLAYLIST: SELECT DJ .............................................................................................................................68
FIGURE 87 - CREATE PLAYLIST: SELECT RAPPER ......................................................................................................................68
FIGURE 88 - PLAYLIST CREATION: SELECT MOOD ....................................................................................................................69
FIGURE 89 - PLAYLIST CREATION: OUTPUT PLAYLIST ...............................................................................................................70
FIGURE 90 - PLAYLIST CREATION: SAVE PLAYLIST ....................................................................................................................70
FIGURE 91 - PLAYLIST CREATION: SAVE PLAYLIST PAGE ............................................................................................................71
FIGURE 92 - PLAYLIST CREATION: YOUTUBE MUSIC PLAYLIST ...................................................................................................71
FIGURE 93 - PLAYLIST CREATE: YOUTUBE MUSIC SUGGESTIONS ................................................................................................72
FIGURE 94 - PLAYLIST CREATE: SPOTIFY PLAYLIST ...................................................................................................................72
FIGURE 95 - PLAYLIST CREATION: ARTIST RECOMMENDATION BASED ON ARTIST CHOICE ...............................................................73
FIGURE 96 - PLAYLIST CREATION: ARTIST RECOMMENDATION BASED ON MOOD CHOICE ...............................................................73
FIGURE 97 - PLAYLIST CREATION: RECOMMENDED ARTIST PLAYLIST ...........................................................................................74
FIGURE 98 - MOODIFY TESTING CHECKLIST ...........................................................................................................................75
FIGURE 99 - JAKOB NIELSON'S 10 HEURISTICS © LANGMAJER, M. (2019) .................................................................................76
FIGURE 100 - HEURISTIC 1, PICTURE 1 .................................................................................................................................77
FIGURE 101 - ALBUM SUGGESTIONS BASED ON MOOD ...........................................................................................................77

Page 6 of 119
FIGURE 102 - SAVING PLAYLIST OPTIONS ..............................................................................................................................78
FIGURE 103 - ISSUE IN AESTHETIC AND MINIMALISTIC DESIGN ..................................................................................................78
FIGURE 104 - RECTIFIED ISSUE FOR AESTHETIC & MINIMALISTIC DESIGN ....................................................................................79
FIGURE 105 - RECOGNITION RATHER THAN RECALL.................................................................................................................79
FIGURE 106 - SUS QUESTIONNAIRE RESPONSES ....................................................................................................................81
FIGURE 107 - USER GUIDE PAGE 1 ......................................................................................................................................82
FIGURE 108 - USER GUIDE PAGE 2 ......................................................................................................................................83
FIGURE 109 - USER GUIDE PAGE 3 ......................................................................................................................................84
FIGURE 110 - USER GUIDE PAGE 4 ......................................................................................................................................85
FIGURE 111 - USER GUIDE PAGE 5 ......................................................................................................................................86
FIGURE 112 - USER GUIDE PAGE 6 ......................................................................................................................................87
FIGURE 113 - USER GUIDE PAGE 7 ......................................................................................................................................88
FIGURE 114 - USER GUIDE PAGE 8 ......................................................................................................................................89
FIGURE 115 - USER GUIDE PAGE 9 ......................................................................................................................................90
FIGURE 116 - USER GUIDE PAGE 10 ....................................................................................................................................91
FIGURE 117 - USER GUIDE PAGE 11 ....................................................................................................................................92
FIGURE 118 - USER GUIDE PAGE 12 ....................................................................................................................................93
FIGURE 119 - USER GUIDE PAGE 13 ....................................................................................................................................94

CHAPTER 1 - INTRODUCTION

1.1 INTRODUCTION TO THE PROJECT

Page 7 of 119
Every person's life is brightened by music. Music has the ability to excite the intellect. There are several
aspects of music to which one might listen and pay attention. One can be aware of the themes or
melodies, the harmony, the driving or calm paces, the richness of the notes, the activity of a piece, how
the sounds are formed, or how they all interact with one another, all while attempting to deduce how the
composer envisioned the work.1111` (-Galindo, G. 2003)

There has been a seismic shift in the way music is delivered since the upheaval brought about by digital
streaming services. With a varied selection of streaming channels and millions of listeners just a few
clicks away, a digital release has more options than ever to reach the right audience. Playlists are an
essential component of any successful music distribution plan. They're the current counterpart of a
mixtape: a compilation of songs curated by someone to portray a mood, topic, or sentiment. These
playlists are usually created by editorial teams on music platforms like Deezer and Spotify or by users on
the app. (-Calvi, J. 2016)

Gone are the days when individuals thoughtfully picked songs to listen to and recorded them in a specific
order on a device. Very few individuals go to the trouble of selecting music and compiling them into a
playlist. When listening to music, consumers are more likely to skip a song than choose one. Hence,
playlisting makes their lives easier. (-Mamtani, V. 2019)

Playlists have never been more important for both music distributors and the audiences they hope to
reach. Because of the evolving nature of how music is discovered and absorbed, playlists have become
one of the most efficient strategies for music discovery and marketing. Many listeners don't want to sit
down and go through record after record until they find their next musical fixation, especially now when
billions of songs are available at the press of a button. Today's average customer prefers to jump fast
between songs that are new, interesting, and trending. They want to find music that has been specifically
picked for their preferences, and they want it to be presented in an easily accessible style. Playlists allow
you to precisely do that.

According to Statista, figures from Nielsen's 2019 Mid-Year Music Report demonstrate how vital
streaming has become to the music business during the few years since its inception, even though
growth is not as robust as it once was. As compared to physical albums, digital albums, and digital track
sales, on-demand streaming (playlists) is consumed way more.

Figure 1 - Music Consumption Statistics

According to a study, the essential qualities of music streaming services are sound quality and pricing,
with 56 percent of respondents selecting both of these as the most important factors people consider
while listening to music online. Consumers also demand convenience, with 48 percent indicating that the

Page 8 of 119
most significant element of service was the simplicity of constructing a playlist, 43 percent saying the
same about downloading music, and 36 percent prioritizing the ease of discovering new songs and
artists. Less significant were brand trust, video availability, and the ability to use a virtual assistant
provided by a business efficiently.

Figure 2 - Most Important Features of Streaming Services

The main goal of this project is to help people listen to music based on the current mood they are feeling.
Often people go through many different moods throughout the day and probably want to listen to a song
based on the current mood they are in. The created website, Moodify asks the user to select what mood
they are currently feeling (Happy, Sad, Angry, In Love, Heartbroken, etc.). Additionally, they can select
the genre they want to listen to and an artist of their choice. Based on user input, the website
recommends them a playlist to listen to, which would be very time efficient and help their mood.

Furthermore, based on the artist selection, the website also recommends other artists to listen to in order
for users to discover new artists and music that they may potentially like. This would benefit both the user
and the artist.

This report provides a detailed description of how the project was carried out. It starts with listing the
project objectives, deliverables, and problem definition, followed by a literature review to analyze similar
existing systems. The work requirements and data analysis are laid out in detail as well. After the data
analysis, the report structure describes the analysis, design, and implementation of the website. It also
includes pictures of the early design wireframes to depict what changed while the final design was being
created. Front-end JavaScript tests and evaluations were also carried out.

1.2 PROJECT OBJECTIVE

Conducting user research and identifying the importance of playlists.

Reviewing already existing user recommended created playlists and learning how they do so.

Conducting a survey to analyze users' favorite artists, genres, and moods in order to recommend
people an artist to listen to based on the genre and favorite artists they pick.

Analyzing survey data on Tableau to find trends and patterns.

Page 9 of 119
1.3 PROJECT AIM
The aim of the project is to analyze user behavior and mood when it comes to music and offer them a
platform (website) where they can select the mood, genre, and artist of choice and receive a personalized
playlist that they can listen to in order to help their mood.

1.4 PROJECT DELIVERABLES

Survey Data has been collected from Google Forms.

Visual Analysis, which was done on the survey data collected. This will be done on Tableau, and
interesting patterns and trends will be discussed with screenshots.

The website code of Moodify, which has the front-end and back-end of the website allowing users
to create a playlist for themselves based on their mood.

Final Report, which includes the discussion of the topic, the interesting facts and findings, the
analysis of the survey data, the literature review and discussion of the website.

Entity Relationship Diagram that demonstrates the links between entity sets recorded in a
database.

A use case diagram, which explains the specifics of the system's users (also known as actors)
and their interactions with it.

1.5 PROBLEMS ADDRESSED

Creating playlists can be a laborious task. Many aspects, such as transitions between tracks or
the musical or thematic coherence of the selected musical pieces, can theoretically be addressed
in the composition process. In general, a huge number of choices can quickly overwhelm people,
leading to issues such as "choice overload" or "over choice”. Although consumers only know a
subset of the available possibilities when creating playlists, they still find it difficult to recall
acceptable tracks from memory, and they frequently seek to expand their music collection by
discovering new tunes. (-Hagen 2015)

It is very difficult and time-consuming for users to sit and decide on a playlist when they would
most likely be busy with their daily tasks. With the help of Moodify, they can auto-generate a
playlist instead of manually doing it.

When users are in a bad mood, they might not feel like searching for new music or creating a
playlist by browsing through catalogs. The website will make it more time-efficient and easy for
them to generate a playlist.

This website would also make it an easy way for users to find their interests and discover new
music/artists that they would like. It would provide a great user experience for them by making the
content more personalized for each user.

CHAPTER 2 - LITERATURE REVIEW

Page 10 of 119
The literature review's goal is to obtain an understanding of the current research and debates pertinent to
a certain topic or area of study and to deliver that information in the form of a written report. In this case,
the research is based on applications and software that create auto-generated music playlists for users
and the data they collect to create algorithms. (-Western Sydney University Library, 2016.)

2.1 PLAYLISTING INTRODUCTION

A playlist is available in most music/audio playing software applications and digital devices. A playlist is a
collection of music or audio files that are either pre-installed or manually added by the user. (-Hope, C.
2020.)

Playlists are incredibly significant today, both for users and for music specialists. Playlists are an excellent
way for users to discover new music and artists. From the standpoint of the researcher/specialist, it is
critical to recognize that music is absorbed through hearing, and playlists formalize that listening
experience. Playlists, like musicians, tracks, and albums, are unit components that may be found and
suggested. (-Papreja, P. 2019)

Some programs distinguish between auto-generated playlists with music placed in a certain sequence
and playlists with tracks mixed. However, the nomenclature used in the literature is not always constant.
(-Logan, 2004). Hence in this review, the focus is more on playlists that are auto-generated and not
manually selected by the user to be created.

2.2 PECULIARITY OF MUSIC RECOMMENDATIONS

The major aspects of music recommendation would include-

The size of most commercial music catalogs is in the tens of millions of music pieces. As a result,
scalability has become a far more essential issue in music recommendations.

Recommending the same piece of music at a later date may be welcomed by the user of an
MRS.

Songs are frequently consumed in a passive, background manner. While this is not an issue in
and of itself, it can have an impact on preference elicitation. When employing implicit feedback to
infer user choices, the fact that a user is not paying attention to the music (and so not skipping a
song) may be misinterpreted as a good signal. (-Zamani Hamed, 2018.)

Music is known to elicit powerful emotions. However, it is a two-way road since consumers'
emotions influence their musical selections. Because of the close association between music and
emotions, the difficulty of automatically characterizing music in terms of emotion words is a huge
debated academic topic known as music emotion recognition (MER). (-Deldjoo, Y. 2018)

Contextual or situational factors have a significant impact on music taste, usage, and
engagement behavior. For example, a listener will most likely build a different playlist when
preparing for a long road trip with friends than when getting ready to go to work with a colleague
on a Monday morning. Location (e.g., listening at work, while commuting, or resting at home) is
one of the most commonly studied categories of context (-Schedl, M. 2018).

Page 11 of 119
2.3 AUTOMATED PLAYLIST GENERATION
The majority of research on automated playlist production is conducted in two areas: Recommender
Systems (RS) and Music Information Retrieval (MIR).

2.3.1 RECOMMENDER SYSTEMS

Recommender Systems are a type of application that give consumers customized suggestions (-Burke,
2002). The playlist generation problem may be thought of as a subset of the recommendation problem,
namely the recommendation of collections (-Hansen and Golbeck, 2009). Additionally, recommender
systems may be used to propose other musical resources such as albums, performers, or performances,
and playlist creation can be thought of as a subset of music (track) recommendations. However, the
peculiarity of playlist creation is that the list's attributes and track orderings should be significant. It should
have consistency in the form of genre and feels.

Famous music applications like Spotify use Recommender Systems to create their auto-generated
playlists. They developed an API called” Spotify Developer API” to download all the playlisting data and
retrieve song information when a new song is uploaded on the streaming app. They clean the raw data by
removing noise which includes fake songs, duplicate songs, any outliers, etc. Once the data is cleaned,
they annotate the cleaned data for genre information (Pop, Ballad, R&B, Rap, etc.). Following that,
Spotify does an evaluation check on the embedded data based on their task criteria and finally builds a
recommendation engine with the analyzed playlist embeddings along with retrieving search results by
making use of the closest neighborhood approach.

Figure 3 - Music Recommender System

©Papreja, P. Building Music Playlist Recommendation System. 2019

Because of the rapid surge in playlist usage, playlist suggestion is critical to music services today.
However, from a research standpoint, playlist suggestion has become akin to playlist
prediction/creation and continuation rather than playlist discovery in recent years. Playlist discovery, on
the other hand, is an important aspect of the entire playlist recommendation pipeline since it is an
effective approach to help users find existing playlists on the site.

Page 12 of 119
Most auto-generated playlists that are created by recommender systems are done by sequence-to-
sequence learning (Seq2Seq). Sequence-to-sequence learning is the process of training models that
translate sequences from one domain to sequences from another domain (-Chollet, F. 2017). The main
reason for using sequence-to-sequence learning is that playlists could be viewed as sentences and tracks
as words in a sentence. Sequence-to-sequence learning has been widely utilized in recent years to train
good phrase embeddings in applications such as neural machine translation. Developers could employ
the connection playlist:songs::sentences:words and draw inspiration from natural language processing
study to structure playlist embeddings in the same manner that sentences are embedded.

Finally, unlike the usual e-commerce situations studied in RS research (Jannach et al. 2012), repeated
consumption of things is highly prevalent in the song domain and audio logs such as those accessible
from last.fm demonstrate that many users listen to their favorite tunes over and over.

2.3.2 MUSIC INFORMATION RETRIEVAL

Music Information Retrieval (MIR) is focused on the extraction, analysis, and application of information
about any type of musical element (for example, a song or a music artist) at any depiction level (for
example, audio signal or name of a music artist.) Music information retrieval (MIR) is the multidisciplinary
study of obtaining information from music. MIR is a small but expanding field of study with several real-
world applications (-Schedl M. 2003).

Given its apparent economic appeal, most media content owners and distributors (e.g., Philips, Sony,
Apple) are actively interested in research on the subject, while many libraries are attempting to include
MIR support into their online digital services.

Simple MIR systems retrieve data based on a text query entered by the user, such as 'We could be
heroes.' In certain circumstances, the text is compared to the text data linked with albums and tracks,
making the system virtually identical to any text-based search engine (e.g., Google, Yahoo). However,
given the features of the material being retrieved, systems that can accept "musical" inquiries such as
compositions, sung melodies (query by humming), or recorded audio segments are required (query by
example). This idea addresses the latter scenario (-Cleverdon, CW. 1997).

Figure 4 - MIR Process

© Joe, K. Multi Version Music Search using Acoustic Feature Union. 2009

Page 13 of 119
The purpose of querying by example is to get musical pieces from a huge library of digital music content
based on their resemblance to an example audio document. The capacity to inquire by example is a
critical prerequisite for MIR systems. It presents a number of hurdles, such as computational and
complexity concerns, the construction of a suitable testbed, and the selection of an appropriate
representation for the audio in the query and audio collection. The system's ability to identify similarities is
determined by the sound representation used. Existing techniques for representation selection may be
generally classified as those seeking to find low-level (acoustic) similarities and those attempting to
quantify high-level (e.g., note, melody, etc.) similarities.

In general, playlist generation might be regarded as a retrieval job if its definition is confined to selecting
music that fulfills a user query. However, the quality of a playlist cannot be determined just by its
individual pieces because there may be desirable traits that apply to the list as a whole, such as diversity,
the presence of things unfamiliar to the user, or seamless track transitions. Moreover, explicit queries are
only one type of user input. Other critical subjects in MIR include automated feature extraction, track
classification, and similarity analysis. These kinds of automatically detected data may definitely be used
to develop a playlist creation system (-Dopler, et al. 2008).

MIR, on the other hand, has the opportune benefit of multiple current programs to construct massive
testbeds of audio content for access and analysis by the scientific community. Furthermore, a developer
called Downie has created a testbed containing almost three terabytes of songs that can be accessed
remotely without violating copyright. The systems are often built with open-source C++ and Matlab
libraries (-Downie, J.S. 2003).

2.4 MOST COMMONLY USED PLAYLIST RECOMMENDATION TECHNIQUES

Based on the research above, it would be more ideal to create an automated playlist generation based on
a recommender system than a music information retrieval system. Some of the reasons include-

Music Information Retrieval systems are more complex compared to a music recommender
system.

MIR systems are more time-consuming compared to an MRS systems.

Easier to gain access to a larger variety of songs on an MRS system since it can connect to the
Spotify API.

Music Recommender Systems use sequence-to-sequence learning to track songs by sentences

and even words which would help understand the tone, meaning, and emotion behind a particular
song.

Some of the most commonly used playlist recommendation techniques have been analyzed below.

2.4.1 PLAYLIST RECOMMENDATION BASED ON REAL-TIME BEHAVIOR

The quickest music recommendation system is based on the user's real-time activity. This type of musical
recommendation algorithm believes that the user's state will remain consistent in a short period of time
and that the next track the user could listen to will have comparable qualities to the track that the user is

Page 14 of 119
currently listening to. The system makes extensive use of song editing features to propose songs to users
that have the same or comparable authors, song titles, and lyrics. The system then makes suggestions
for users based on the acoustic features of songs and suggests music with similar tempo, groove, and
timbre. The system describes music using mood parameters and suggests songs with comparable mood
qualities to current tracks to users (-Shi, J. 2021).

Most of these projects explain music from a single point of view and then give suggestions to users;
however, these recommendations are frequently biased or one-sided. For example, the system would not
be appropriate for individuals who are not susceptible to acoustic aspects but are susceptible to
emotional features. The system only analyzes text materials relating to songs, using semantic
characteristics to characterize songs, and produces satisfactory results. (-Shi, J. 2021).

2.4.2 PLAYLIST RECOMMENDATION BASED ON LONG-TERM BEHAVIOR

Playlisting recommendation based on long-term user behavior evaluates and assesses all tracks that
individuals have listened to, employing collaborative filtering recommendations, which has lately been the
most common recommendation algorithm. It scans the users' social environments and uses swarm
intelligence to produce suggestions for them. The system describes records using acoustic characteristics
and then computes the mean acoustic descriptors of all tracks listened to by individuals as useful
features. Since musical recommendation based on users' long-term behavior includes a thorough
examination of consumers, this type of algorithm may frequently achieve a good recommendation impact.
Applications such as the newly launched YouTube Music follow this recommendation system. (-Shi, J.
2021).

2.5 APPLE MUSIC VS SPOTIFY

Apple Music creates its playlist generation with the help of Music Information Retrieval (MIR) whilst
Spotify creates theirs with the help of Recommender Systems (RS).

Apple Music contains over 60 million songs, so customers are guaranteed to find music from artists they
know and love, as well as musicians they haven't heard of yet. Furthermore, Apple Music is well-known
for striking exclusive partnerships with artists such as Drake and Taylor Swift in exchange for early
content releases. This is due in large part to Apple's premium service; after the three-month free trial,
consumers are obliged to have a paid membership. While Apple Music has more material, Spotify's music
collection is still vast, with over 50 million songs and an additional 40,000 uploaded every day. Unlike
Apple Music, Spotify has a podcast section, with over 700,000 now available. Spotify has been
broadening its material catalog over the years in order to become the perfect music streaming service. (-
Maranan, H. 2021).

Page 15 of 119
 Figure 5 – Spotify Personalized Playlists

© Spotify, 2021. Auto-generated playlists by Spotify based on the User’s listening history.

Customers are familiar with Spotify's ability to propose music based on your listening behavior. Sure,
Apple Music's human-curated channels help consumers discover new songs, but they pale in comparison
to Spotify's recommendation algorithms. Users get recommendations for new music based on what's in
their playlist at the bottom of each category. Users also receive personalized playlists like Discover
Weekly, which provides hours of new music for them to discover depending on their listening
habits. Moreover, Spotify also contains a selection of hand-picked playlists with the most recent songs,
which are automatically rearranged for each user based on their preferences.

Figure 6 - Spotify Playlist based on user history

© Spotify, 2021. Curated based on User’s listening history on Spotify

Spotify-created playlists are personalized based on the user's listening patterns (what you enjoy, save,
skip and share) as well as the listening habits of people who share the same preferences as the user.
Spotify has five different kinds of personalized playlists for its users.

Page 16 of 119
 Figure 7 - Types of Personalized Playlists curated by Spotify

© Spotify’s Personalized Playlists, made on Draw.io

Apple Music's newest version has a separate 'Listen Now' tab with song recommendations based on the
user's listening history. ' Listen Now replaces the former 'For You' page's unwieldy design with a far more
simplified layout that has a striking resemblance to the Spotify app's main page. (-Maranan, H. 2021).

Apple delights in its playlists, using humans as curators rather than depending only on algorithms. The
service's playlists are excellent for discovering new music since they cover a wide range of genres and
are created by artists and professional editorial teams. However, since it is created by humans, there
would be a slight bias in the songs being selected for the playlist (Kulm, Z. 2020).

Figure 8 - Common Playlists by Apple Music

© Apple Music, 2021. Playlists curated by Apple Music

After careful analysis and research on Apple Music and Spotify, the scoreboard can conclude the overall
satisfaction of the applications.

Page 17 of 119
 Figure 9 - Apple Music vs Spotify

© Spotify vs. Apple Music Scoreboard. Created on Draw.io

Overall, Spotify leads in terms of having better playlist generation as well as being more personalized to
each and every user on their platform. They are able to do this with deep learning and machine learning
algorithms in their systems to recommend perfectly tailored playlists to users.

2.6 TARGET CHARACTERISTICS FOR MUSIC PLAYLIST GENERATION

To build playlists, three inputs are required: a pool of tunes, a prior knowledge database, and certain
target attributes. Background knowledge is any information that may be utilized to judge if a playlist and
its tracks meet the goal requirements. Typically, information is divided into four categories: audio signal,
meta-data, Social Web data, and usage data.

Audio material is an apparent sort of data that may be employed in the playlist construction process.
Many MIR researchers aim to automatically extract musical elements from audio signals, such as pitch,
loudness (-Blum et al. 1999), rhythm, chord changes (-Tzanetakis 2002), or simply compact
representations of the frequency spectrum (-Pohle et al. 2005).

The advantage of relying on the audio signal versus, say, expert comments is that it is more accurate.

To automate the playlist generating process, the required target attributes must be supplied in a machine-
interpretable format. These traits connect to the playlist's overarching purpose or intended aims. There
are three sorts of such information that may be supplied into the automatic creation process: (1)
contextual sensor data, (2) explicit preferences and constraints, and (3) previous user choices.

Users' previous preferences can be clearly indicated (ratings), but the algorithm can also guess them.
Estimates are often derived from the user's own track collection on the device; the user's listening history,
or playlists that the user has already built.

The user's track library might be viewed as an indicator of the user's long-term taste. If the collection is
large, it is uncertain if the user has ever listened or even knows all of the tunes (-Lamere 2012). As a

Page 18 of 119
result, this type of information might be untrustworthy and lacks "negative" feedback. As a result, the
user's listening logs are frequently more instructive, especially when "skip" actions are logged.

2.7 DIFFICULTIES FACED IN AUTOMATED PLAYLIST GENERATION

A variety of factors make automatic playlist creation difficult. Firstly, even if the process is automated, the
huge number of possible tracks that can potentially be included in the playlists – already, personal music
libraries can be enormous (-Dias and Fonseca 2010) – and the corresponding immense amount of extra
data that must be processed can result in high computational complexity. While certain calculations may
be done in an asynchronous model-learning process, the challenge is that many application situations
need "on-demand" playlist production or quick response to user feedback.

Secondly, listeners' tastes and circumstances can have a significant impact on the perceived
appropriateness or performance of a playlist, and two users may experience the same set list differently
even if listened to in the same situation. Thus, the problem is not only to select tracks that match the
characteristics of the present user's taste and situation, but also to accurately identify the user's
preference and circumstance in the first place.

Lastly, a major problem is gathering information about the songs. Extracting features from an audio file for
a large number of tracks might take a long time to compute. However, if we rely on user-provided
annotations, acquiring a collection of high-quality annotations for each track might be a significant
challenge. Often, just a portion of the recordings are annotated (-Celma, 2010), and the quality of the
annotations might be low and subjective.

2.8 RESEARCH METHODOLOGY OR TECHNIQUES

For primary data, conducting surveys would be vital to gather quantitative data to understand the basic
characteristics of a large number of people.

For secondary data, researching on already existing websites/applications that create auto recommended
playlists and learning how about their efficiency, functionality, advantages, and disadvantages would
provide qualitative and quantitative data (case studies and literature review).

For this paper, research was conducted through different articles online and Google scholar's case
studies that helped in the research. Few quantitative data was found on websites that had conducted
surveys and interviews in countries.

2.9 RESEARCH METHODS

When it comes to data collection and analysis, quantitative research is concerned with numbers and
figures, whereas qualitative research is concerned with words and meanings. Both are necessary for
acquiring various types of information.

Numbers and graphs are used to express quantitative research. It is used to validate or test hypotheses
and assumptions. This research method can be used to develop generalizable facts about a subject.
Tests, observations recorded as numbers, and surveys with closed-ended questions are examples of
common quantitative procedures.

Words are used to express qualitative research. It is utilized to comprehend ideas, thoughts, or
experiences. This form of study allows you to gain in-depth knowledge about issues that are not generally

Page 19 of 119
known. Interviews with open-ended inquiries, observations reported in words, and literature studies that
investigate concepts and theories are all common qualitative approaches.

QUANTITATIVE RESEARCH QUALITATIVE RESEARCH

The emphasis is on putting theories and Exploration of ideas and formulation of a theory or
hypotheses to the test. hypothesis.
Math and statistical analysis were used to analyze The data was analyzed by summarizing,
the data. classifying, and analyzing it.
Mostly conveyed by tables, numbers, and graphs. Mostly conveyed in words.
Requires a lot of respondents Requires a few respondents.
Closed (multiple choice) questions. Open-Ended Questions.
Testing, measurement, objectivity, and Understanding, context, complexity, and
replicability are keywords. subjectivity are keywords.
Usually conducted if people want to confirm or Usually conducted if people want to learn more
test anything. about anything.
2.10 RESEARCH FINDINGS AND ANALYSIS (RESULTS)
Using Spotify's technique of creating auto-generated playlists using recommender systems proves to be
more time-efficient and easier as compared to using Apple Music’s technique with MIR (Music Information
Retrieval). Using a recommender system will allow the user to receive a more personalized playlist
catered to the mood they select.

Recommender Systems will allow the cleansing of data, adding of new data, removing unwanted data
and more, keeping the database fresh and clean.

CHAPTER 3 - DATA ANALYSIS

3.1 INTRODUCTION TO THE DATA BEING COLLECTED

In order to gain a better understanding of users’ music preferences in terms of genres, artists, and
moods, a survey was conducted via Google Forms so that the website would be able to be as accurate
as possible when a user wants a recommended playlist. It would help in understanding which artists to
keep on the website, what genres to add on the website and what are the most common moods people
usually go through. There was a total of 21 questions based on the following topics:

Various moods.
Music taste.
Favorite artists and genres.
How they consume music.
How their mood affects their music taste.

3.2 DATA COLLECTION PROCESS

Page 20 of 119
A consent form was sent out to each individual, which they had to sign if they chose to do the survey. The
survey was sent to people via different social media, including Messenger, Instagram, WhatsApp, and
even Gmail.

The only data collected for analysis is from the survey responses. The questions have been attached in
the appendix. Survey responses reached up to 48. The data was then converted into an excel sheet and
loaded on Tableau for further analysis.

If some insufficient data was noticed in some sections, then online research was conducted to resolve
plot holes. For example, research was conducted to identify the most commonly listened to artists when
people are angry. The data was then converted into an excel sheet and loaded on Tableau for further
analysis. The redundant data or inaccurate data was cleaned and removed. Some examples of data that
was cleaned included-

Data where a pop artist’s name was written instead of a rap artist’s name, was removed.
Data where users did not fill in three names and left it blank, were also removed.
Case sensitivity was fixed. (Example – Ariana Grande and ariana grande counted as two different
people, so it needed to be changed to the same case.

3.3 RESEARCH OBJECTIVES

Identifying which music platform individuals use the most while listening to music.
Identifying an individual’s favorite genre to listen to.
Learning how often people listen to music on a weekly basis.
Understanding and analyzing how a person’s mood can affect their genre choice and artist
choice.
Learning users’ top three favorite artists in different genres and analyzing if there is any pattern
between the artists that have been mentioned.

3.4 INITIAL KEY FINDINGS

Basic data were analyzed first. Out of all the survey responses, 74.47% of users listen to music daily. The
research shows that the majority of the people listen to music frequently, with only a handful of 4.26%
listening three times a week.

Page 21 of 119
 Figure 10 - Music Listening Pattern

As for music lifting up moods for users, 74.47% claimed that listening to music puts them in a better
mood, whilst 25.53% chose ‘sometimes’. Not a single user selected ‘No’ (0%).

Figure 11 - Mood Lift Up by Music

Considering users listening patterns, in terms of music genres, 40.43% of users chose ‘Pop’ music as
their favorite genre. Other genres that some users chose were ‘Hip-Hop/Rap’ (19.15%) and
‘Alternative/Indie’ (8.64%). Hence on the ‘Moodify’ website, Pop artists have more moods as compared to
other genres.

Most users listen to music on Spotify (17 responses), while some listen on their phone’s music app and
YouTube Music.

Page 22 of 119
Based on music preference, 85.11% of users prefer listening to a mix of old-generation music, and new-
generation music whilst 10.64% only prefer the latter.

Figure 12 - Genre, Music App and Preference

When users were asked about the most common mood they face, 38% of the users (17 responses)
claimed that they feel neutral (neither sad nor happy). 23.4% feel happy/excited most of the time. All of
the other responses were 6.38% of the total.

Figure 13 - Most Common Mood

Users were asked to select their three most favorite pop artists. The survey responses show that majority
of the people chose ‘Ariana Grande’ (18 users) and ‘Taylor Swift’ (17 users). The male artist that was
named the most was ‘Ed Sheeran’ (7 Users).

Page 23 of 119
https://towardsdatascience.com/create-automatic-playlists-by-using-deep-learning-to-listen-to-the-
music-b72836c24ce2

https://uxdesign.cc/how-could-spotifys-auto-generated-playlists-be-improved-a-ux-analysis-
ed9dbd2fc17f

https://www.gilbertgalindo.com/importanceofmusic#:~:text=Music%20can%20raise%20someone's%20
mood,we%20experience%20in%20our%20lives.&text=It%20is%20an%20important%20part,an%20urge
%20to%20create%20music.

https://imusician.pro/en/resources/guides/playlists-building-audiences

https://uxdesign.cc/the-playlist-theory-98d4d83402d0

Western Sydney University Library. (2016). Essay purpose. Retrieved from

https://westernsydney.edu.au/studysmart/home/assignment_help/writing

https://www.computerhope.com/jargon/p/playlist.htm

https://towardsdatascience.com/building-music-playlists-recommendation-system-564a3e63ef64

Figure 14 - Favorite Pop Artists

As for users’ favorite rap artists, there were 9 responses stating that they did not like rap music; hence
their survey data related to rap music was considered an outlier and cleaned. Based on the survey
responses, most people named ‘Eminem’ as one of their favorite rappers (14 users). The female rapper
that was named the most were ‘Nicki Minaj’ and ‘Cardi B’, both selected by 11 users.

Figure 15 - Favorite Rap Artists

Considering users’ favorite bands, majority of them have listed ‘Maroon 5’ (24 Users) as one of their

Page 24 of 119
Considering users’ favorite bands, majority of them have listed ‘Maroon 5’ (24 Users) as one of their
favorites. ‘One Direction’ was selected by 17 users, while girl group ‘Little Mix’ garnered a total of 15.

Figure 16 - Favorite Bands

As for users’ favorite DJ, majority of responses had ‘Avicii’ mentioned as one of their top 3 (21 Users). Not
far behind were ‘Marshmello’ and ‘Martin Garrix’ with 17 mentions each. Out of all the different categories
(Pop, Hip-Hop, Bands and Dance), Avicii was the artists with the most mentions.

The purpose of analyzing users’ favorite pop and hip-hop artists as well as their favorite bands and DJs
are to be aware of which artists/bands/DJs to keep on the top of the list while creating the playlist-
recommendation website. Keeping the most popular artists on the top of the list while help the user to
identify them faster, making it more time efficient and useful.

Figure 17 - Favorite DJs

Page 25 of 119
3.5 ANALYSIS OF MOOD WITH GENRES
When users are in an angry mood, 28.26% of users prefer listening to calm alternative/pop songs.
21.74% of users claimed they would listen to Rock/Metal songs, and 19.57% chose upbeat
pop/empowerment songs. In this case, if a user selects ‘angry’ as their mood on the playlist generation
website, the website suggests them an artist that makes music under this genre.

Figure 18 - Genre When Angry

When users are in a happy mood, 51.06% of them prefer listening to upbeat pop songs, and 27.66% of
users like to listen to hip-hop/rap music. Hence when a user creates a playlist, the website would
additionally suggest to the user an artist that makes upbeat pop songs to listen to.

Figure 19 - Genre When Happy

Page 26 of 119
When users are in a sad mood, 57.45% of users prefer listening to ballads whilst 23.40% prefer calm
songs. Hence the website suggests users’ artists that make more ballad songs if they select their mood
as ‘Sad’.

Figure 20 - Genre When Sad

3.6 ANALYSIS OF MOOD WITH ARTISTS

When users are in a happy mood, 4 responses said that the artist they find themselves listening to the
most was ‘Taylor Swift’, more than any other artist selected. Hence, when a user creates a playlist on
‘Moodify’ and selects their mood as happy, once the playlist is generated, it would also suggest them to
listen to ‘Taylor Swift’s Happy Music Playlist’.

Figure 21 - Most Frequently Listened to Artist (Happy Mood)

Page 27 of 119
Additionally, when users are in a sad mood, 5 responses claimed that the artist they listen to the most
was once again, ‘Taylor Swift’. 3 responses also named ‘Adele’ as their most listened to artist when they
are feeling low. Hence, when a user creates a playlist on ‘Moodify’ and selects their mood as sad, once
the playlist is generated, it would also suggest them to listen to ‘Taylor Swift’s Sad Music Playlist’ as well
as ‘Adele’s Sad Music Playlist’.

Figure 22 - Most Frequently Listened to Artist (Sad Mood)

3.7 ANALYSIS OF ARTISTS

The main reason why users were asked to select their top 3 artists on the survey was to try to
recommend close-to-accurate suggestions while users selected their artists of choice on the playlist
generating website. (For example, if a user selects ‘Taylor Swift’ as their artist of choice and any mood
after the playlist is generated, it suggests other artists to listen to who may be making music similar to
Taylor Swift.)

This would be beneficial to both the user and the artist. The user will be able to expand their music taste
and discover artists and songs they made have never heard before. The artist will gain new streams and
gear up a larger fanbase.

The following were interesting patterns found in the survey responses.

3.7.1 Pop Artist Recommendation

All users that selected ‘Alicia Keys’ as one of their top pop artists also selected the Weekend. The
number ‘1’ inside the circle denotes the number of times a survey response had the same three names
listed in the same order together.

Page 28 of 119
 Figure 23 - Alicia Keys

Similar patterns were noticed with other artists as well. ‘Shawn Mendes’ and ‘Taylor Swift’ appeared
together a few times even though the first artist selected was different. They appeared together in 3
different responses.

Figure 24 - Shawn Mendes & Taylor Swift

Whenever a user selected ‘Katy Perry’ they also chose ‘Ariana Grande’.

Figure 25 - Katy Perry & Ariana Grande

When ‘Dua Lipa’ was selected, 6 survey responses listed out ‘Ariana Grande’ and ‘Taylor Swift’ as their
other 2 favorite pop artists. This was the most common sequence out of all the survey responses. Hence,

Page 29 of 119
when a user selects ‘Dua Lipa’ as an artist of choice, after the playlist is generated, the website
recommends them to also check out playlists based on ‘Ariana Grande’ and ‘Taylor Swift’. This would
work vice versa as well.

‘Taylor Swift’ appeared a total of 8 times when ‘Dua Lipa’ was selected, and ‘Ariana Grande’ was listed 7
times.

Figure 26 - Dua Lipa

When ’Ed Sheeran’ was selected, 5 out of 6 also chose ‘Bruno Mars’ in their top 3. ‘Ariana Grande’ was
chosen twice. Hence it would be best to suggest ‘Bruno Mars’ and ‘Ariana Grande’ when a user selects
‘Ed Sheeran’ and vice versa.

Figure 27 - Ed Sheeran

Pop Artist Recommendation Summary

The following patterns were noticed and would be suggested on the Moodify website for artist
suggestions.

Alicia Keys  The Weeknd

Shawn Mendes  Taylor Swift

Katy Perry  Ariana Grande

Dua Lipa  Ariana Grande  Taylor Swift

Ed Sheeran  Bruno Mars  Ariana Grande

Page 30 of 119
3.7.2 Hip-Hop/Rap Artist Recommendation

Whenever ‘Cardi B’ was selected as one of the users’ favorite rappers, 3 users also selected ‘Doja Cat’
and ‘Megan Thee Stallion’ as their two other favorite rappers. ‘Megan Thee Stallion’ appeared a total of 5
times on the list. ‘Eminem’ appeared 4 times but in different combinations.
Hence the best suggestions would be ‘Doja Cat’, ‘Megan Thee Stallion’ and ‘Eminem’.

Figure 28 - Cardi B

When a user listed ‘Drake’ as one of their top 3 rappers, 4 of 7 also chose ‘Eminem’ as their favorite.
‘Kendrick Lamar’ was listed thrice and two of the users who listed him also listen ‘Kendrick Lamar’.

The best suggestions would be ‘Travis Scott’, ‘Eminem’ and ‘Kendrick Lamar’.

Figure 29 – Drake

When a user listed ‘Nicki Minaj’ as one of their favorite rappers, ‘Doja Cat’ and ‘Eminem’ were selected
together twice. ‘Drake’ and ‘Eminem’ were also selected together twice.

‘Doja Cat’ was listed a total of three times whilst ‘Eminem’ was listed 4 times. Hence the best suggestions
would be ‘Eminem’ and ‘Doja Cat’.

Page 31 of 119
 Figure 30 - Nicki Minaj

Hip-Hop/Rap Artist Recommendation Summary

The following patterns were noticed and would be suggested on the Moodify website for artist
suggestions.

Cardi B  Doja Cat  Megan Thee Stallion  Eminem

Drake  Travis Scott  Kendrick Lamar  Eminem

Nicki Minaj  Doja Cat  Eminem  Drake

Eminem would be a common suggestion for rap artists since his name is listed in majority of the survey
responses as one of the users’ favorite rappers.

3.7.3 DJ Recommendation

When a user selected ‘Alan Walker’ as one of their top DJ’s, 3 of 6 users also selected ‘Marshmello’.
‘Calvin Harris’ was also selected thrice. Hence the two would be a good suggestion when ‘Alan Walker’
is selected and vice versa.

Figure 31 - Alan Walker

‘Avicii’ was the most selected DJ by users. When ‘Avicii’ was selected as the one of the users favorite
DJs, 3 responses also had ‘Calvin Harris’ and ‘Martin Garrix’ along with him. 3 responses even had
‘Marshmello’ and ‘Martin Garrix’ along with ‘Avicii’.

‘Martin Garrix’ was listed a total of 8 times with ‘Avicii’. ‘Marshmello’ was listed a total of 7 times. Hence
the best DJ suggestion would be ‘Martin Garrix’ and ‘Marshmello’.

Page 32 of 119
 Figure 32 – Avicii

Whichever users selected ‘DJ Khaled’ as their favorite DJ also selected ‘Jonas Blue’ (2 of 2).

Figure 33 - DJ Khaled

When users selected ‘DJ Snake’ as their favorite DJ, 4 of 6 users also selected ‘Marshmello’. ‘Martin
Garrix’ was selected thrice.

Hence it would be ideal to suggest ‘Martin Garrix’ and ‘Marshmello’ to the users that choose ‘DJ Snake’
as their DJ for the playlist on Moodify.

Figure 34 - DJ Snake

DJ Recommendation Summary

The following patterns were noticed and would be suggested on the Moodify website for DJ suggestions.

Alan Walker  Calvin Harris  Marshmello

Avicii  Martin Garrix  Marshmello

DJ Khaled  Jonas Blue

Page 33 of 119
DJ Snake  Martin Garrix  Marshmello

3.7.4 BAND RECOMMENDATION

When users selected ‘5 Seconds of Summer’ as their favorite artists, 6 of 10 also chose ‘Maroon 5’. ‘Little
Mix’ was chosen 4 times from which 2 were selected along with ‘Maroon 5’.

Hence the best suggestion for ‘5 Seconds of Summer’ would be ‘Maroon 5’ and ‘Little Mix’.

Figure 35 - 5 Seconds of Summer

When ‘BTS’ was chosen, ‘Maroon 5’ and ‘Imagine Dragons’ appeared twice proving to be the best
suggestion for ‘BTS’.

Figure 36 – BTS

When ‘One Direction’ was listed by a user, 3 users selected ‘Coldplay’ and ‘Maroon 5’ as their 2nd and 3rd
favorite band. 2 users selected ‘Fifth Harmony’ and ‘Little Mix’ as their 2nd and 3rd favorite band while 2
other users selected ‘BLACKPINK’ and ‘Fifth Harmony’. ‘Little Mix’ however, appeared in 5 different
combinations.

Hence when ‘One Direction’ is chosen, the best suggestion would be ‘Maroon 5’, ‘Coldplay’ and ‘Little
Mix’.

Page 34 of 119
 Figure 37 - One Direction

Whenever ‘One Republic’ was listed, ‘Maroon 5’ was listed as well (2 of 2). Similarly, whenever ‘Twenty-
One Pilots’ was listed, ‘Panic! At The Disco’ was listed too (2 of 2).

Figure 38 - One Republic and Twenty-One Pilots

Band Recommendation Summary

The following patterns were noticed and would be suggested on the Moodify website for band
suggestions.

5 Seconds of Summer  Maroon 5  Little Mix

BTS  Maroon 5  Imagine Dragons

One Direction  Coldplay  Maroon 5  Little Mix

One Republic  Maroon 5

Twenty-One Pilots  Panic! At The Disco

X Ambassadors  Maroon 5

Maroon 5 would be a common suggestion for bands since their name is listed in majority of the survey
responses as one of the users’ favorite bands.

3.8 DATA LIMITATIONS

Page 35 of 119
Although data collection and analysis were done with the survey data, some of the data required for the
website were not able to be collected from the survey responses. Some of these included-

Commonly listened to artists when a user is in an angry, relaxing, or heartbroken mood.

The genre of music a user prefers listening to when heartbroken or feeling relaxed.

Some of the artists were not selected with another artist multiple times, making it difficult to
decide which other artist to suggest for the user to listen to.

However, these issues were resolved by researching online and identifying common genres for specific
moods as well as artist suggestions for the artists that had insufficient data in the survey responses.

CHAPTER 4 - REQUIREMENTS SPECIFICATION

4.1 TOOLS & REQUIREMENT TECHNIQUES

The creation of this project required a lot of tools and techniques, which are listed below.

Having access to the internet for research purposes.

Canva was used to create the logo for the website.
Visual Code to code the front end and back end of the website and add the algorithms required
for playlist generation. This would be necessary for creating the login and register of the user as
well, with the help of JavaScript to store the data on the session and local storage.
Excel was used to load all the survey data.
Tableau was used to analyze all the primary data received from the survey responses.
draw.io was used to create the use case diagram and the entity-relationship diagram.
Using Miro to create the wireframes for the website.
Selenium IDE will be used to test the front-end functionality of the website.

4.2 FUNCTIONAL REQUIREMENTS

Functional requirements revolve around what a product/system must perform, as well as its features and
functions. They are components or functions that programmers must build in order for users to complete
their objectives. In general, functional requirements explain how a system behaves under certain
situations (-Alex, 2018). The functional requirements for creating the Moodify website include-

The website must allow a user to navigate through the different tabs of the website.
The website must allow a user to click buttons which directs them to the necessary section of the
system.
The website must allow a user to successfully create an account as well as log in with the same
credentials.
The website must allow the user to discover different artists on the platform and navigate through
their individual details.

Page 36 of 119
 The website must allow the user to create a playlist by allowing them to start the process, select
their desired genre from the available options, select their desired artist from the available options
and select their mood from the available options.
The website must allow the user to view each song when they click “Hear Song” on the playlist
created as well as allow them to view ‘artists recommendation’ based on the mood selected and
artists selected mentioned below the created playlist.
The website must allow the user to save the playlist by allowing them to save it as an image,
saving the playlist on YouTube music, or saving the playlist on Spotify.

4.3 NON-FUNCTIONAL REQUIREMENTS

Non-Functional Requirements are not connected to website/system functionality but rather specify how
the system should work (-Alex, 2018). The requirements for the Moodify website include-

The website should not lag and must load quickly depending on the number of users using it at
the same time.
The website must be simple and easy to understand for the users by keeping the design
aesthetic and simple and not too clustered so that the user does not get confused.
Ensuring a user’s data is secure by keeping a login/register tab on the website.
Ensuring that the website is compatible with most devices.
Ensuring that the number of critical failures when testing the website is less than none.
Making sure that the user would find it easy to navigate through the features of the website and
accomplish their goals.

4.4 USER INTERFACE REQUIREMENTS

The user interface is critical to the usefulness of a software or website. Content display, application
navigation, and user help should all be included in the application/website (-IBM, 2021).

The website created should be usable without reading a printed guide. Users should find the
interface simple to understand and easy to navigate through.
Users should find the website functional and time-efficient. There should not be any bugs or
delays on the website.
Users should be able to accomplish their desired tasks on the website in the expected average
time and the fewest number of clicks.
Users should not feel lost on the website or confused about where to go in order to achieve a
specific task.

4.5 SOFTWARE REQUIREMENTS

A software requirements specification (SRS) is a document that outlines what the program will accomplish
and how it should behave (- Lane, C. 2021).

The website will allow the users to view the new music released each week, new albums
released, and upcoming music.
The website will allow the users to discover their favorite artists and gain more knowledge about
them.
The website will allow the users to create their own personalized playlist by selecting their genre,
artist, and mood.

Page 37 of 119
 The website will allow the users to save their playlist on Spotify or YouTube Music.
The website will display recommended artists to listen to based on the user’s mood selected and
the artist selected.

CHAPTER 5 - ANALYSIS AND DESIGN FOR THE WEBSITE

5.1 USE CASE DIAGRAM

Use case diagrams are used to collect a system's needs, including internal and external impacts. The
majority of these criteria are design requirements (-Tuna, 2017). This use case diagram helps understand
the way the website functions along with the inputs it receives and outputs it displays.

Figure 39 - Use Case Diagram

Page 38 of 119
5.2 E/R DIAGRAM
An entity relationship diagram (ERD) depicts the relationships between entity sets in a database. In this
sense, an entity is an object, a data component (-SmartDraw, 2020).

Figure 40 - Entity Relationship Diagram

5.3 ACTIVITY DIAGRAM

An activity diagram is essentially a diagram or flowchart that depicts the transition from one stage to
another. The action can be defined as a system operation. The control flow is directed from one task to
the next. This flow might be sequential, branching, or concurrent in nature. Activity diagrams deal with all
types of flow control by employing various features such as fork, join, and so on (-Tuna, F. 2019).

Page 39 of 119
Figure 41 - Activity Diagram

Page 40 of 119
5.4 FEATURES OF THE WEBSITE (STORYBOARD)
A storyboard is a diagram that depicts the screens that comprise a system's user interface and how a
user can navigate between the various functions. Before developers dig in and start constructing,
storyboards allow them as programmers and creators to visualize and optimize this part of their system (-
Chadwick, R. 2021).

This storyboard highlights all the main important features of the Moodify website. It would cover the
following-

Creating an account and signing in.

Viewing newly released music (weekly basis), newly released albums, and upcoming releases.
Learning the importance of music, the benefits of auto-generated playlists, and how music affects
mood and vice versa.
Creating a playlist most suited for the user by allowing them to select their genre, mood, and artist
of choice.
Allowing the user to save the playlist on Spotify, YouTube music or as an image for future use.
Recommending the user with artists to listen to (based on their artist of choice and current mood).

Page 41 of 119
 Figure 42 - Moodify Storyboard

5.5 WIREFRAMES OF INITIAL DESIGN STAGES

A wireframe is a schematic or blueprint that may be used to plan and describe the structure of software or
a website that is being developed (-Guilizzoni, 2018).

Before starting the coding process of the website, it was important to first create a skeleton design of the
website to further understand the important features it should have and what the whole layout would look
like.

In the early stages, it was decided to have four different sections-

Home page
Login page
Make a playlist page
About Us page

Page 42 of 119
 The wireframes were designed on a website called ‘Miro’. The user would be introduced to the
website with a brief explanation of what the website’s features are. It was also important to have a
section called ‘New Music this week’ which would help the user discover new music and hopefully
their next favorite jam.

Figure 43 - Initial Home Page Design

Having a page to register and create an account would be necessary to save playlists that the user would
create.

Page 43 of 119
 Figure 44 - Initial Register Design Wireframe

Once the user registers, they would log in through the sign in page.

Figure 45 - Initial Sign in Wireframe

Page 44 of 119
The ‘about us’ page would focus more on theory and learning and explains to the user the importance of
music, how it affects people’s moods and the benefits of auto-generating a playlist.

Figure 46 - Initial About Us Design

In the ‘make a playlist’ page, the system would first ask the user to select a genre to influence the playlist.

Figure 47 - Initial Genre Selection Design

Page 45 of 119
After selecting the genre, the system would continue to ask the user to select their artist of choice to
influence the playlist.

Figure 48 - Initial Artist of Choice Design

Lastly, the system would ask the user to select their current mood that would influence the playlist.

Figure 49 - Initial Mood Choice Design

Page 46 of 119
After the user clicks ‘next’, the system would receive all the data and process it, after which the system
would output the created playlist. Based on the user’s picks of mood and artist, the system also
recommends artists and songs for the user to listen to.

Figure 50 - Created User Playlist Initial Design

5.6 WEBSITE IMPLEMENTATION & SOURCE CODE

The website implementation was a long and tedious process. Coding was done on “Visual Studio Code”
and occasionally on “Brackets”. A folder titled ‘Moodify’ was made, and the process of coding began with
the home page. To get some inspiration for the layout and color scheme, a few YouTube tutorials were
watched to get inspired of how dynamic a website can look. The logo for the website was created on
Canva. Spotify and YouTube Music were used for playlist generation. (Screenshots of the Webpage
design can be found in the next section – “Website Functionality and Final Design”)

The home page was designed with the help of HTML, CSS, and JavaScript. Each of these was stored as
separate files and linked to one another using the required linking tags.

Page 47 of 119
 Figure 51 - moodify.html

Adjacent to the ‘music quote of the day’ is a play button that will play a calm bird chirping noise if the user
clicks on it. The reason behind adding this was simply because the website is based on music and mood
and the best way to incorporate the message on the home page is by adding calm and peaceful music to
soothe the user. The functionality of starting and stopping the song was done with the help of a
JavaScript function.

Figure 52 - Opening Song JavaScript Function

Page 48 of 119
 The opening page of the website also includes a trailer of what to expect on the website along with a
remix of all the popular artists that may be featured on the website. This was accomplished by attaching
the video to HTML and using JavaScript functions and CSS to make it look appealing.

The coding on the home page included four sections-

New Music this Week

New Albums Released
Upcoming Music
Footer

All buttons on the website have a hover effect with the help of CSS ‘: hover’ effect.

The next page that was coded was the “Learn More” page. In the initial design, it was called “About Us”
but it wasn’t really about the website. It was more about the importance of music and mood as well as the
benefits of auto-playlist generation. Hence, the title was renamed.

The “Login/Register” page was also coded with the help of HTML, CSS & JavaScript. The details entered
by the user were stored on the session local storage. The function ‘store’ stores all the details inputted by
the user whilst also ensuring that the password meets the guidelines.

Figure 53 - Function Store

The images for the website were either taken from Google Images or created with the help of Canva.

The ‘Discover’ page is slightly different from the rest of the pages. It focuses more on gaining knowledge
about the artists that are used on the Moodify website. The JavaScript function with the help of CSS,
helps to make the moon, stars, and mountains in the background image move as the user scroll down the
page. The background image was found separately on a website and was stored as an .eps file.

Page 49 of 119
 Figure 54 - Moving Background

JavaScript onclick functions were used to link the artist page to the discover page.

The ‘Make a Playlist’ page featured a slideshow container on the page which was created with the help of
CSS and JavaScript. The slideshow was automated to switch every 2.6 seconds with the help of auto
play delay function.

Figure 55 - Auto play Slideshow

When the user clicks “Create a Playlist”, it takes them to the main Playlist Creation page. Content is
hidden with the help of CSS and JavaScript using the ‘Class List Hide’ function. When the user clicks

Page 50 of 119
“start” the first section of the content is unhidden, where the user is asked to select a genre for the
playlist.

The same procedure is followed when the user selects a genre. The genre selected is inputted into the
system, and the container is hidden. The artist container is then unhid for the user to select an artist to
influence the playlist. The artist chosen is also inputted into the system. Lastly, the user is asked to select
their current mood. The mood chosen is also inputted into the system.

With the help of JavaScript, based on the values of genre, artist, and mood, it outputs the desired playlist
for the user.

Figure 56 - Function to Start the Playlist

Page 51 of 119
 Figure 57 - Html for Playlist Creation

The images for the mood options, genre options, and artist options were found on Google Images.
However, the artist picture after the playlist is displayed was edited on the app ‘Photogrid’ and used.
Online research was conducted to identify the mood of different songs by the artist chosen, and lyrics
were read through on the platform ‘Genius’ to ensure that the lyrics match with the mood and are best
suited for the user.

For some artists, interviews of theirs were watched where they explained songs that they have created in
order to get more accurate data about their music and recommend those particular songs on Moodify
based on the mood best fitted for the song.

When the playlist is outputted to the user, it also suggests them with a few artists they make like based on
their artist choice and also based on their mood choices. These suggestions were implemented with the
help of the survey analysis that was conducted in the earlier stages of the project. The survey analysis
helped identify which artists were selected together, and suggestions were based on that. A few artists
that were not chosen together proved to be a limitation; hence for those artists, online research was
conducted to identify similar artists. The process for suggesting artists based on mood choice was the
same.

Artists based on artist choice, had a ‘discover’ button that linked the page to their YouTube channel,
where the user could discover their music.

Artists based on mood choice had a ‘view playlist’ button which, when clicked, redirected them to a
playlist centered around that artist but for the same mood.

Page 52 of 119
 Figure 58 - Artist Suggestions

When the user clicks on the “Save Playlist” button, it redirects them to a page where they can save the
playlist in 3 different ways. They can either save the playlist as an image, as a playlist on Spotify or on
YouTube Music since those were the most commonly used music streaming platforms based on the
survey conducted in the earlier phase of the project.

In order to achieve this, the HTML file was linked with Spotify and YouTube Music.

Figure 59 - Saving Playlist Code

Page 53 of 119
In conclusion, the different tools used for developing the website were-

Visual Studio Code & Brackets to code the website.

Google Chrome to run the HTML file.
HTML, JavaScript, and CSS to make the website functional.
Google Images for various pictures, videos, and GIFs required for the website.
Canva and PhotoGrid to edit some of the pictures.
VSDC Video Editor to create a trailer for the website (available on the Homepage).
Google to research various sections that were going to be on the website.
Wikipedia to get information about the artists and their music.
Genius to read song lyrics and identify the genre, feel and mood of the music.
YouTube to link the songs with the videos.
YouTube Music and Spotify to link the created playlist on Moodify with them.
Apple Store and Apple Music were used to view track listing of the albums and view all of the
artists' discography.
Microsoft Teams and Zoom to conduct Usability testing for Moodify.
Selenium IDE to test the runtime of the website.

5.7 WEBSITE FUNCTIONALITY & FINAL DESIGN

5.7.1 FUNCTIONALITY OF THE WEBSITE

The ease with which a person may navigate around a website, find the information they need, and/or
achieve their purpose on the website is described as website functionality. Websites that prioritize utility
above appearance or elegance provide greater benefits to consumers (-Misiak, I. 2020).

A good web design centers on the demands and issues of the main user. Testing, methodology, key
demographic, information architecture, and web design are also important considerations. Users are on
the page because they need something as fast as possible. The website must be able to match the
expectations of the site's visitors (-Misiak, I. 2020).

Moodify’s functionality is found throughout all the sections of the website. Buttons direct the user to the
desired locations of the website, a navigation bar to direct the user from one section to another, links that
redirect the user to the desired page, and playlisting creation, which allows the user to create a
personalized playlist based on their mood and artist of choice.

5.7.2 WEBSITE FINAL DESIGN

Moodify’s opening page (home page) features a music quote of the day on the left and the tagline on the
right. The play button allows you to have a bird chirping calm noise running in the background as the user
navigates through the website. The song can be paused by clicking the pause button.

The navigation bar allows the user to navigate through the five different sections of the website- Home,
Login, Make a Playlist, Discover and Learn More.

Page 54 of 119
 Figure 60 - Moodify Home Page #1

There even is an ‘Explore Now’ button which when clicked, scrolls down to the next section on the
homepage which is a trailer video of all the features that are available on the website. It allows the user to
be aware of the different things to expect as they navigate through the site.

Figure 61 - Trailer Video on Homepage

The next section in the home page is called ‘New Music This Week’ which updates every week. It
features a gif of the song, the title, the artist, the genre and also the mood of the song. This is because
the website’s main objective is to intertwine music and mood. Hence, the mood is displayed for each song
in the form of emojis so that the user can check out the new tracks based on whichever mood fits them at
that moment. It also includes a “Hear Song” button which when clicked, directs the user to the music
video of that song on YouTube.

Page 55 of 119
 Figure 62 - Moodify: New Music This Week

The next section on the homepage is “New Albums Released” which features all the albums released by
artists in the past few weeks. It displays the cover art, title, artist, mood, and genre of the album. The
mood defines the mood of the album as a whole, as described by the artists themselves in the interviews
they did to promote the album. It also has a link to hear the album, which directs the user to YouTube
Music, where they can listen to the album in its full entirety.

Figure 63 - Moodify: New Albums Released

The next section is the “Coming Soon” section which features all the artists who have announced music
that will be coming out in the near future. This is just to keep the users informed if any of their favorite
artists have announced any music that will be coming soon. Some songs have the date mentioned, while
some songs just have the month since the date has not been announced yet.

Page 56 of 119
 Figure 64 - Moodify: Coming Soon

The last section of the home page is the footer.

Figure 65 - Moodify Footer

The “Login” Page features a form for the users to fill to register an account on moodify if they are new to
the site. It asks the user the enter their Email, Username and Password. The password has a few
guidelines that need to be kept in mind while creating an account. This is for the safety of the users.

Page 57 of 119
 Figure 66 - Moodify: Register

If the user already has an account or has completed the registration, they can click the ‘log in here’ button
that is underneath the form.

The button redirects them to the login form where they have to enter their email, username, and
password. Their details have been stored on the session local storage on HTML and confirms if the data
entered matches. If it does, it tells the user that they have logged in successfully.

Figure 67 - Moodify: Log In

Page 58 of 119
The “Learn More” page features a brief description of the importance of music as the header along with
the author.

Figure 68 - Moodify: Learn More

As the user clicks the “See More” button it scrolls down to explain to the user about how music can affect
people’s mood as well as the Benefits of auto-playlist generation which can be done in the “Make a
Playlist” page.

Figure 69 - Moodify: Learn More #2

Benefits of auto-playlist generation which can be done in the “Make a Playlist” page.

Page 59 of 119
The “Discover” section is very different from the rest of the website. This section is where users can
discover all the artists that are featured on Moodify. The artists were selected based on the survey that
was conducted in the earlier phases on the project. Analysis was performed to identify the most
commonly listened to artists and those artists were selected to be on Moodify.

The discover page allows the user to learn more about the artists and their music.

The discover page uses JavaScript functions and an .eps file to make the background move. Therefore,
when the page is first clicked, only the “Explore” button is visible. But as the user scrolls down, the title
“Discover Artists” slowly appears, the moon starts to set, and the stars move.

Figure 70 - Discover Page when Opened

Figure 71 - Discover Page while Scrolling

The artists are divided into two different sections. The Pop Artists featured and the Hip-Hop artists that
are featured. Some of the artists such as ‘Olivia Rodrigo’ and ‘Katy Perry’ are not directly featured to

Page 60 of 119
select as an artist of choice while creating a playlist but are suggested to listen to when another artist of
choice is selected.

Figure 72 - Discover Pop Artists

Figure 73 - Discover Hip Hop/Rap Artists

When an artist is clicked, for example ‘Taylor Swift’ a new tab is opened which features details about
Taylor Swift.

It features a picture of the artist, the artist’s name, the years they have been active in the music industry,
their age and little bit about them. These details were found through websites such as Wikipedia and
Britannica.

Page 61 of 119
 Figure 74 - Detailed Artist Discover Page

As the user scrolls further down, they can also view all the albums released by the artist. Their latest
album is displayed first and then it moves in the order from their most recent album to their least recent
album.

Figure 75 - Discover: Albums Released

Additionally, since Moodify also focuses on the mood for the music, it suggests the user which albums to
listen to, based on the mood they are experiencing. The suggestion was done based on online research
about the overall mood of the album as well as how the artist described the albums during interviews they
did.

Page 62 of 119
 Figure 76 - Discover Album based on Mood Suggestions

The next section features all the singles that have been released by the artist. This does not mean all the
songs that have been released by an artist. A single is a song that is being released before the release of
an album, usually in order to promote the album and give the audience a little taste of what the album
would sound like. It can also be released after the album release by sending the song to radio, doing live
performances of the song, making a music video for the song, and promoting it on talk shows and more (-
Mark, V. 2019).

The singles are placed in order from most recent to least recent. It also displays the album the single is
from and the year it was released.

Figure 77 - Discover: Singles Released

Page 63 of 119
Lastly, the “Latest Interview” and “Latest Performance” section allows the user to view the artists’ latest
interview and performance with the help of a button that links it to the video on YouTube.

Figure 78 - Discover: Latest Interview & Performance

This format is the same for all the other artists that are available on the “Discover” page.

Figure 79 - Discover: Katy Perry

The next page features the “Make a Playlist” section where there is an image slideshow running on auto
play to describe the steps to make a mood-based playlist. The steps include-

Page 64 of 119
 Select a music genre
Select an artist of choice
Select a mood
Receive your playlist

The image slideshow was made with the help of JavaScript functions.

Figure 80 - Make a Playlist: Image Slideshow

As the user scrolls down, there is a button that allows them to start the process of making a playlist.

Figure 81 - Making a Playlist

Once the button is clicked, the process begins. The user is asked if they want to start the process or exit
to the home page.

Page 65 of 119
 Figure 82 - Create Playlist: Step 1

When the user clicks the start button, the system then asks the user to select a genre from the 4 following
options- Pop, Bands, Dance, or Hip-Hop.

Figure 83 - Create Playlist: Select Genre

When the user selects pop, the system takes in the genre and displays the pop artists to choose from.
There are 6 artists to choose from- Justin Bieber, Ariana Grande, Dua Lipa, Shawn Mendes, Camila
Cabello, or Taylor Swift. The user also can also go back to the genre and change it if they’d like.

Page 66 of 119
 Figure 84 - Create Playlist: Select Pop Artist

When the user selects bands, the system takes in the genre and displays the bands to choose from.
There are 6 bands to choose from- Maroon 5, One Direction (1D), Little Mix, 5 Seconds of Summer
(5SOS), BTS, or Coldplay.

Figure 85 - Create Playlist: Select Band

When the user selects dance, the system takes in the genre and displays the DJs to choose from. There
are 6 DJs to choose from- Zedd, Calvin Harris, Marshmello, Avicii, Alan Walker, or DJ Khaled.

Page 67 of 119
 Figure 86 - Create Playlist: Select DJ

When the user selects Hip-Hop music, the system takes in the genre and displays the rappers to choose
from. There are 6 rappers to choose from- Cardi B, Eminem, Nicki Minaj, Doja Cat, Megan Thee Stallion
or Drake.

Figure 87 - Create Playlist: Select Rapper

Considering the user wants to make a pop playlist, and clicks ‘Ariana Grande’ as the artist of choice. The
system takes in the value of the artist selected (‘Ariana Grande’ in this case) and the user is then asked to

Page 68 of 119
select what mood they are currently feeling. The process is the same for any artist selected amongst the
4 genres. There are 6 different options to choose from –

Happy, Excited and/or Cheerful

Sad, Depressed and/or Frustrated
Heartbroken
In Love and/or Flirty
Angry, Upset and/or Annoyed
Relaxing, Calm and/or Peaceful

Figure 88 - Playlist Creation: Select Mood

Whichever option is clicked, that value is taken in by the system. For example, if “Happy/Excited/Cheerful”
is selected, the system takes in that value.

The system then runs JavaScript functions and identifies that the genre chosen was ‘Pop’, the artist
chosen was ‘Ariana Grande’ and the mood selected was ‘Happy/Excited/Cheerful’. Based on that, it
generates a Happy Ariana Grande Playlist to the user as an output.

Page 69 of 119
 Figure 89 - Playlist Creation: Output Playlist

The playlist is outputted in the form of a table which includes the album artwork, song name, artist name
(with the artists featured- if any), album name, year released and a song link which links the song to
YouTube.

As the user scrolls down, they have an option to save the playlist as well.

Figure 90 - Playlist Creation: Save Playlist

When the user clicks this button, they are directed to a new tab where they have three different ways to
save the playlist.

Page 70 of 119
The first is to save the playlist on YouTube Music. The second is to save the playlist on Spotify and the
other way is to save the playlist as an image. YouTube Music and Spotify were selected as the two
music streaming platforms to save the playlist on since the research that was conducted in the earlier
phase of the project showed that these were the two most commonly used music streaming platforms.

Additionally, in the literature review phase it was identified that both YouTube music and Spotify use
Recommender System algorithms to create their playlist and their playlists are more widely used and
personalized to each user as compared to other music streaming platforms like Apple Music and
Samsung Music.

Figure 91 - Playlist Creation: Save Playlist Page

When the user clicks the first link, it takes them to the YouTube music link where the playlist has been
created.

Figure 92 - Playlist Creation: YouTube Music Playlist

Page 71 of 119
As the user scrolls down, YouTube music also provides them with more suggestions that they can add to
the playlist as well as related playlists which they can also check out.

Figure 93 - Playlist Create: YouTube Music Suggestions

If the users click the second link on the ‘Save Playlist’ page, it directs them to their created playlist on
Spotify.

Figure 94 - Playlist Create: Spotify Playlist

The playlist on Spotify has all the Ariana Grande songs that are happy but also includes happy songs by
other artists that people listen to, who also listen to Ariana Grande. This was done through Spotify’s
algorithms as well as from the research conducted in the earlier phase.

When the user created their playlist, as they scroll down on that same page, Moodify also recommends
them with artists to listen to based on their mood selected and based on their artist choice.

Page 72 of 119
 Figure 95 - Playlist Creation: Artist Recommendation based on Artist Choice

Figure 96 - Playlist Creation: Artist Recommendation based on Mood Choice

For the artist recommendation based on artist choice, when the ‘Discover’ button is clicked, it directs the
user to the official YouTube Channel of that artist where they can discover their music, songs, videos,
performances and more.

For the artist recommendation based on mood choice, when the ‘View Playlist’ is clicked, Moodify takes in
the value of artist selected and creates a playlist with the new ‘Artist of Choice’. For example, if Katy
Perry’s ‘View Playlist’ button is clicked, Moodify will display the newly Recommended Katy Perry Playlist
for the same mood selected.

Page 73 of 119
 Figure 97 - Playlist Creation: Recommended Artist Playlist

CHAPTER 6 - TESTING AND EVALUATION

6.1 TESTING CHECKLIST
In order to ensure that everything on Moodify is working correctly and doing the task it is assigned to do,
a checklist was created. Additionally, while users perform the usability testing, this checklist will be used
to analyze whether all these points were completed successfully or not.

Page 74 of 119
 Figure 98 - Moodify Testing Checklist

6.2 FUNCTIONALITY TESTING

To test the functionality of the front-end JavaScript on Moodify and the navigation along with checking the
run-time of the website, the ‘Selenium IDE’ tool was used. The Selenium IDE (Integrated Development
Environment) is basically a run/record tool used by a test case developer to create various Selenium Test
cases. It enables the test case developer to autonomously record their test cases depending on browser
interactions. It also allows developers more flexibility in conducting test scenarios. The test developer can
either execute the complete test suite, which consists of numerous test cases or can run a single test
case (-Unadkat, J. 2021).

To use Selenium IDE, a chrome extension was installed to run the software. Tests were done for the
Homepage, the Discover Page, and the Make a Playlist Page. All of the tests were successful, and the
interaction time was acceptable. (Tests Results can be found in Appendix 3)

6.3 HEURISTIC EVALUATION

Heuristic evaluation is a technique in which specialists utilize rules of thumb to test and comment on the
usability of user interfaces in separate guidelines. Evaluators employ well-established heuristics (e.g.,
Nielsen- Molich's) to uncover insights that can assist design teams in improving product usability from the
start (-Kahneman, D. 2016).

Page 75 of 119
 Figure 99 - Jakob Nielson's 10 Heuristics © Langmajer, M. (2019)

It would be extremely difficult to allow all the ten heuristics to shine equally on the website. Hence, only a
few of the heuristics were taken into consideration while the system was being created.

• Match between System and the Real World.

• User Control and Freedom
• Flexibility and Efficiency of Use
• Aesthetic and Minimalistic Design
• Recognition rather than Recall.

6.3.1 PROCEDURE OF HEURISTIC TESTING

Two students from the User Interface/User Experience Design class were reached out to via the social
media platform WhatsApp to ask if they would like to participate in a Heuristic Evaluation Test for the
Moodify website. They agreed to participate, and consent forms were sent out to them which they had to
sign after reading the terms and conditions.

The heuristic testing could not be done face to face due to the ongoing COVID-19 pandemic; hence it
was done online via Microsoft Teams. A date was scheduled for the meeting and the allotted time was
about 20-30 minutes for each participant. The participants were given a list of the heuristics that were
followed and were asked to give feedback on each of the heuristics. They were given access to control
the shared screen so that they could navigate through the website by themselves.

6.3.2 RESULTS OF HEURISTIC TESTING

Match between System and the Real World

No issues were found in this section. The participants claimed that the way you would find albums in the
store would be presented in a similar manner. It would display the album cover, artist name, and genre
just like depicted on the website.

Page 76 of 119
 Figure 100 - Heuristic 1, Picture 1

User Control and Freedom

There were no issues found here by the participants as well. They exclaimed that Moodify allows you to
freely navigate throughout the platform and lets you discover all the artists that are present on the website
and receive album suggestions based on mood.

Figure 101 - Album Suggestions Based on Mood

They also said that the website freely allows you to create a playlist from the 4 mood options, the 24
available artist options and the 6 mood choices.

They found navigating through the website to be simply and easy to understand.

Flexibility and Efficiency of Use

Once again, there were no issues found here. The participants feedback summarized that the user could
have full control over the website when they are using it. There were no restrictions whatsoever. Users

Page 77 of 119
also have the option to download their created playlist in 3 different ways and save it in two different
streaming platforms (Spotify and YouTube Music).

Figure 102 - Saving Playlist Options

Aesthetic and Minimalistic Design

Only one issue was found here which was in the playlist section where the users have to select an artist
of choice. They said that all the buttons were placed at the bottom of the picture except for the ‘Dua Lipa’
button. This made it feel a little inconsistent.

Figure 103 - Issue in Aesthetic and Minimalistic Design

This issue was then analyzed, and it was identified that if the text in the button was more than a certain
number of characters, it would move to the next line. It is originally supposed to be displayed adjacent to
the picture. Hence this was rectified by keeping only the first names of the artists.

Page 78 of 119
 Figure 104 - Rectified Issue for Aesthetic & Minimalistic Design

Recognition Rather than Recall

There were no issues found here by the participants. They liked the fact that the playlist that is created
displays the album artwork and the album name along with the year released. They explained that
sometimes if a user only sees the song and artist name, they might not be able to remember the song
unless they hear it. However, they have a higher chance of recognizing the song if they also have
additional information about it.

Figure 105 - Recognition Rather than Recall

Page 79 of 119
6.4 USABILITY TESTING
Usability testing is a method of determining how simple a system/website is to use by evaluating it with
real people. The users attempt to complete tasks, generally while being observed by the researcher, in
order to determine where they face difficulty and confusion. If additional people have similar concerns,
suggestions will be made to address these usability issues. The tests are conducted with actual users to
determine how 'accessible' or 'practical' a website is and how simple it is for users to achieve their
objectives. (-McRobbie, D. 2019).

6.4.1 PROCEDURE OF USABILITY TESTING

The usability testing was required to be done by participants who are part of the target audience, which
are university students, young adults, or teenagers. Participants were reached out to on different social
media platforms, including WhatsApp, Instagram, and Facebook. Only the participants that had
participated in the first survey in the earlier research phase of the project were selected for the usability
testing since they already had an idea about what the project was.

About 6 participants were recruited to conduct the usability testing. The testing would take place online
via Microsoft Teams or Zoom as per their preference. A consent form was sent out to the participants
before the test. Once signed, the test was conducted online.

The website was presented to the participant via screen share, and they were allowed remote access to
the screen to navigate around the website. They were asked to use the think-aloud protocol to express
their feedback and feelings toward the website. Only three main tasks were given for them to do. Those
three tasks were-

Create an account on Moodify and Log in.

Discover and gain information about any of your favorite artists from the website.
Create a playlist based on your mood today.

6.4.2 ANALYSIS ON BEHAVIOR AND FEEDBACK

All 6 participants were able to do the three tasks successfully. Most of their feedback was positive. The
users were all able to do the tasks in the average time expected. Participants claimed that they loved the
design and aesthetic of the website and found it to be very informative. They found it easy to navigate
through the different features. The method of creating a playlist was very time efficient, according to them.

4 out of 6 users loved the ‘Discover’ Page the most, whilst two users loved the ‘Home Page’ the most. 4
participants even expressed how much they loved the trailer of the website.

When asked for any feedback or suggestions, one user claimed that K-Pop was their favorite genre, and
it would be nice if Moodify implemented K-Pop as a genre in the future.

Another participant said that having more artists in the future would also be great for the website.

This feedback was taken into consideration for the future plans of the website.

6.4.3 USER SATISFACTION SURVEY

Page 80 of 119
After the usability testing was complete, the participants were given a link to a Google Forms Satisfaction
Survey to express their overall feedback on Moodify (Questions and individual answer pie charts are
attached in the appendix). This form would help in analyzing the areas the user liked and areas where
improvement could be made. Six tests were conducted; hence there were six responses. The data was
then loaded onto excel for analysis.

Figure 106 - SUS Questionnaire Responses

For all of the satisfaction questions, none of the users were neutral, dissatisfied or extremely dissatisfied
with any of the questions. For the most part, users were extremely satisfied with the design, process of
creating a playlist, the features and user friendliness of Moodify.

83.3% of the responses said that they would use this service ‘Twice or Once a week’ since they would
stick to their created playlist for a few days before going back to the website to make another one.

Overall, 3 participants rated Moodify ‘Excellent’ and 3 rated it ‘Very Good’.

CHAPTER 7 - USER GUIDE

A user guide or manual for a site is a document designed to assist users in utilizing the site. It is a
professional form of writing with a certain format that should be authored by someone who is extensively
familiar with the product. The language used is appropriate for the intended audience, and terminology is
described in detail or reduced to a minimum (-Patel, R. 2021).

Below is the User Guide that was created for Moodify.

7.1 USER GUIDE FOR MOODIFY

Page 81 of 119
Figure 107 - User Guide Page 1

Page 82 of 119
Figure 108 - User Guide Page 2

Page 83 of 119
Figure 109 - User Guide Page 3

Page 84 of 119
Figure 110 - User Guide Page 4

Page 85 of 119
Figure 111 - User Guide Page 5

Page 86 of 119
Figure 112 - User Guide Page 6

Page 87 of 119
Figure 113 - User Guide Page 7

Page 88 of 119
Figure 114 - User Guide Page 8

Page 89 of 119
Figure 115 - User Guide Page 9

Page 90 of 119
Figure 116 - User Guide Page 10

Page 91 of 119
Figure 117 - User Guide Page 11

Page 92 of 119
Figure 118 - User Guide Page 12

Page 93 of 119
Figure 119 - User Guide Page 13

Page 94 of 119
 CHAPTER 8 - CONCLUSIONS
8.1 EVALUATIONS & SUMMARY
Spotify, a popular playlisting and streaming service, had 345 million monthly active users in December
2020. The pandemic led to a 31% increase in monthly active users for Spotify. According to Spotify, this
figure will rise to 427 million by the end of 2022. According to research, many users continue to utilize
streaming services to listen to their favorite songs on repeat and to create playlists of their favorite music
(-Hoda, T. 2021).

During the research process, it was discovered that streaming platforms like Spotify featured multiple
user-recommended playlists based on various genres, artists, and moods, but they did not mix all
the three categories into one playlist. That was selected to be the core focus of Moodify. Recommending
a playlist to the user based on their genre, mood, and artist of choice.

The survey phase assisted in empathizing with the participants, understanding how people engage with
music in their everyday lives, and how music impacts their mood.

Based on research, artist recommendation, artist selection, genre selection and mood selection were
finalized for Moodify, and the website was created with the help of HTML, CSS, and JavaScript.

Before coding, wireframes were designed, and sections of the website were designed to have a brief
understanding of how to go about coding the front-end of the website. Features were finalized before the
coding process began.

After the website was designed, functionality tests were conducted with the help of Selenium IDE.
Usability testing was also performed with a few participants that had also participated in the research
phase in order to gain feedback about the design and playlisting process. Additionally, Heuristic
Evaluations were done on the website to ensure it followed five heuristics that were taken into
consideration while creating Moodify.

Long-term, the aim is that these recommendations will become so important, so vital to how people enjoy
music, that music fans will be unable to live without them. This is especially true in a market where
streaming services are becoming increasingly common (-Willens, 2015).

8.2 LIMITATIONS
There were a few limitations that were present on the website due to the time constraints of finishing the
entire project. These limitations included-

Only 4 genres to choose from. One of the participants wanted K-Pop to be a genre as well to
have a little more inclusivity.
Each genre had six artists to choose from. These 6 artists were chosen based on the survey
participants’ favorite artists.
The discover page only had the artists that were present on the website’s playlisting process.

8.3 RECOMMENDATIONS AND FUTURE PLANS

Page 95 of 119
Since this is just the beginning of Moodify, there are a lot more features that are planned to be added to
the website. Those include-

The addition of more genres into the playlisting process. Genres planned to be included are
R&B/Soul, K-Pop, Alternative, and Gospel.
The addition of more artists in each genre.
The addition of more artists in the Discover section as the number of artists in the playlisting
process grows.
A section called ‘Mood Tracker’ that would track your mood every time you create a playlist with a
specific mood.
Allowing the user to select up to 2 to 3 artists to influence their created playlist instead of having
only one.

CHAPTER 9 - REFERENCES

Page 96 of 119
• Alfocea, J. (2016). Apple Music beats Spotify in customer satisfaction. [online] Available at:
<https://www.actualidadiphone.com/en/apple-music-surpasses-spotify-customer-satisfaction/>
[Accessed 22nd December 2021]

• Archer, J. (2021). Apple Music vs Spotify: Which is the best music app? [online] Available at:
<https://www.tomsguide.com/face-off/apple-music-vs-spotify-which-is-the-best-music-app>
[Accessed 23rd December 2021]

• Balsamiq. (2019). What are wireframes? [online] Available at:

https://balsamiq.com/learn/articles/what-are-wireframes/ [Accessed 6th January 2022]

• Bonnin, G. (2014). Automated Generation of Music Playlists: Survey and Experiments. [online]
Available at: <https://dl.acm.org/doi/abs/10.1145/2652481> [Accessed 15th December 2021]

• Charu, C. (2016). Content-based recommender systems. In Recommender Systems. Springer,

139–166. [Accessed 20th December 2021]

• Chollet, F. (2017). A ten-minute introduction to sequence-to-sequence learning in Keras. [online]

Available at: <https://blog.keras.io/a-ten-minute-introduction-to-sequence-to-sequence-learning-
in-keras.html> [Accessed 17th December 2021]

• Cross, A. (2015). Why Playlists are growing in Importance. [online} Available at:
<https://www.ajournalofmusicalthings.com/why-playlists-are-growing-in-importance/> [Accessed
15th December 2021]

• Dacrema, M., Cremonesi, P. (2021). Recommender Systems Reproducibility and Evaluation.

[online] Available at: <https://www.researchgate.net/project/Recommender-systems-
reproducibility-and-evaluation> [Accessed 27th December 2021]

• Gazzarini, A. (2019). Introduction to Music Information Retrieval. [online] Available at:

<https://www.slideshare.net/AndreaGazzarini/introduction-to-music-information-retrieval>
[Accessed 19th December 2021]

• Kamehkhosh, I., Bonnin, G. (2020). Effects of recommendations on the playlist creation behavior
of users. [online] Available at:
<https://www.researchgate.net/publication/333319405_Effects_of_recommendations_on_the_pla
ylist_creation_behavior_of_users> [Accessed 21st December 2021]

Page 97 of 119
• Knees, P., Schedl, M. (2017). New Paths in Music Recommender Systems Research. [online]
Available at: <http://www.cp.jku.at/tutorials/mrs_recsys_2017/overview_paper.pdf> [Accessed
19th December 2021]

• Kulm, Z. (2020). Best Apple Music Playlists. [online] Available at:

<https://www.soda.com/music/best-apple-music-playlists/> [Accessed 21st December 2021]

• Lamere, P. (2018). Recsys Challenge 2018: automatic music playlist continuation. [online]
Available at: <http://www.recsyschallenge.com/2018/> [Accessed 23rd December 2021]

• Maranan, H. (2021). Apple Music vs Spotify. [online] Available at:

<https://www.soundguys.com/apple-music-vs-spotify-36833/> [Accessed 21st December 2021]

• Pachet, F. (2002). Music Retrieval. [online] Available at:

<http://www.eecs.qmul.ac.uk/legacy/easaier/files/technical/retrieval/musicretrieval.pdf> [Accessed
19th December 2021]

• Schedl, M. (2018). Current Challenges and Visions in Music Recommender Systems Research.
[online] Available at: <https://arxiv.org/pdf/1710.03208.pdf> [Accessed 16th December 2021]

• Scribblr. (2019). Research Methods: How to collect and Analyze Data. [online] Available at:
<https://www.scribbr.com/category/methodology/> [Accessed 2nd January 2022]

• Tuna, F. (2019). Use Case Diagram [online] Available at:

<https://www.tutorialspoint.com/uml/uml_use_case_diagram.htm> [Accessed 7th January 2022]

• Willens, M. (2015). In Streaming Music Wars, Your Playlist Has Become the Weapon of Choice.
[online] Available at: <https://www.ibtimes.com/streaming-music-wars-your-playlist-has-become-
weapon-choice-2218326> [Accessed 27th December 2021]

• Galindo, G. (2003). The Importance of Music in our Society. [online] Available at:
<https://www.gilbertgalindo.com/importanceofmusic> [Accessed 31st October 2021]
• Dwivedi, R. (2020). What are Recommendation Systems in Machine Learning? [online] Available
at: <https://www.analyticssteps.com/blogs/what-are-recommendation-systems-machine-learning>
[Accessed 31st October 2021]

• Kamehkhosh, I. (2019). Effects of Recommendations on the Playlist Creation Behavior of Users.

[online] Available at: <https://link.springer.com/article/10.1007/s11257-019-09237-4#Abs1>
[Accessed 1st November 2021]

Page 98 of 119
• Richter, F. (2019) Infographic: Streaming Dominates Music Consumption in the U.S. (no
date) Statista Infographics. [online] Available at: <https://www.statista.com/chart/10185/music-
consumption-in-the-us/.> [Accessed 18th March 2022]

• MN2S (2019) The Rising Importance of Playlists in Music Distribution [online] Available at:
<https://mn2s.com/news/label-services/rising-importance-of-playlists-in-music-
distribution/#:~:text=Being%20included%20on%20the%20right.> [Accessed 19th March 2022]

• Organ, C., Bottorff, C. (2021) Work Breakdown Structure as A Project Management Tool, Forbes
Advisor. [online] Available at: <https://www.forbes.com/advisor/business/what-is-work-
breakdown-structure/.> [Accessed 3rd April 2022]

• SmartDraw (2019) Entity Relationship Diagram (ERD) - What is an ER Diagram? [online]

Available at: <https://www.smartdraw.com/entity-relationship-diagram/.> [Accessed 3rd April 2022]

• Lucidchart (2019) UML Use Case Diagram Tutorial [online] Available at:
<https://www.lucidchart.com/pages/uml-use-case-diagram.> [Accessed 3rd April 2022]

• Project Manager (2019) Gantt Chart – Long Form Info Page. [online] Available at:
<https://www.projectmanager.com/gantt-chart.> [Accessed 4th April 2022]

• Gotting, M. (2022) Most important features of music streaming services among adults in the U.S.
[online] Available at: <https://www.statista.com/statistics/819865/important-features-music-
streaming-services/.> [Accessed 22nd March 2022]

• Shi, J. (2021) ‘Music Recommendation Algorithm Based on Multidimensional Time-Series Model

Analysis’, Complexity. Edited by Z. Lv, 2021, pp. 1–11. doi:10.1155/2021/5579086. [online]
Available at: <https://www.hindawi.com/journals/complexity/2021/5579086/> [Accessed 23rd
March 2022]

• Altexsoft (2021) Functional and Nonfunctional Requirements: Specification and Types. [online]
Available at: <https://www.altexsoft.com/blog/business/functional-and-non-functional-
requirements-specification-and-types/.> [Accessed 11th April 2022]

• Indeed Team (2021) 9 Nonfunctional Requirements Examples - Indeed Career Guide. [online]
Available at: <https://www.indeed.com/career-advice/career-development/non-functional-
requirements-examples.> [Accessed 12th April 2022]

Page 99 of 119
• Moran, S. (2017) Activity Diagram, UML Diagrams Example: Relationships between Activates
and Business Entities - Visual Paradigm Community Circle. [online] Available at:
<https://circle.visual-paradigm.com/relationships-activates-business-entities/> [Accessed 3rd April
2022]

• Chadwick, R. (2022) Create Powerful Storyboards - Software Design and Development. [online]
Available at: <https://ryanstutorials.net/software-design-and-
development/storyboards.php#:~:text=A%20storyboard%20is%20a%20diagram> [Accessed 9th
April 2022]

• Misiak, I. (2020) What Website Functionality Means in 2020 - Element Three. [online] Available
at: <https://elementthree.com/blog/what-website-functionality-means-in-
2020/#:~:text=Website%20functionality%20is%20defined%20by.> [Accessed 11th April 2022]

• Mark, V. (2021) What Is a Single in Music? (Single vs. EP vs. Album, Cost). [online] Available at:
<https://hiphopmakers.com/what-is-a-single-in-music> [Accessed 12th April 2022]

• DailyRapFacts (2021) What Does ‘Single’ And ‘Album’ Mean? [online] Available at:
<https://dailyrapfacts.com/16632/what-does-single-and-album-mean/> [Accessed 12th April 2022]

• Unadkat, J. (2021) Selenium IDE Tutorial: Definition, Features, Benefits - BrowserStack. [online]
Available at: <https://www.browserstack.com/guide/what-is-selenium-
ide#:~:text=Selenium%20IDE%20(Integrated%20Development%20Environment> [Accessed 14th
April 2022]

• Langmajer, M. (2019) 10 usability heuristics every designer should know, Medium. [online]
Available at: <https://uxdesign.cc/10-usability-heuristics-every-designer-should-know-
129b9779ac53> [Accessed 15th April 2022]

• Interaction Design Foundation (2019) What is Heuristic Evaluation? The Interaction Design
Foundation. [online] Available at: <https://www.interaction-design.org/literature/topics/heuristic-
evaluation.> [Accessed 15th April 2022]

• Waldo, S., Stevenson, G. (2019) Real world user interface examples. [online] Available at:
<https://slideplayer.com/slide/14786047/> [Accessed 12th April 2022]

Page 100 of 119

 • IBM (2019) User interface requirements [online] Available at:
<https://www.ibm.com/docs/en/db2/9.7?topic=considerations-user-interface-requirements>
[Accessed 13th April 2022]

• Kruger, N. (2020) How to Write a Software Requirements Specification (SRS

Document), Perforce Software. [online] Available at: <https://www.perforce.com/blog/alm/how-
write-software-requirements-specification-srs-
document#:~:text=A%20software%20requirements%20specification%20(SRS)%20is%20a%20do
cument%20that%20describes.> [Accessed 14th April 2022]

• Patel, R. (2021) A Complete Guide on How to Create User Manual for Website - The
CloudTutorial. [online] Available at: <https://www.thecloudtutorial.com/how-to-create-user-
manual-for-website/#:~:text=A%20user%20manual%20or%20guide> [Accessed 16th April 2022]

• Hoda, T. (2021) Using Spotify playlists to design consumer experience - linkedin.com. [online]
Available at: <https://www.linkedin.com/pulse/using-spotify-playlists-design-your-consumers-
tasneem-hoda-acsm/?trk=read_related_article-card_title> [Accessed 1st April 2022]

CHAPTER 10 – APPENDIX

TABLE OF CONTENTS
Serial Content Page Number
Number
Appendix 1 Supervisor Meeting Logs 106
Appendix 2 Source Code 106

Page 101 of 119

 Appendix 3 Selenium IDE Testing 107
Appendix 4 Ethical Screening Form 109
Appendix 5 Ethical Approval Form 111
Appendix 6 Research Phase Consent Form 113
Appendix 7 Usability Testing Consent Form 114
Appendix 8 Heuristic Evaluation Consent Form 115
Appendix 9 Moodify User Satisfaction Survey Questions 116
Appendix 10 Moodify User Satisfaction Survey Responses 119

APPENDIX 1 – SUPERVISOR MEETING LOGS

Page 102 of 119

APPENDIX 2 – SOURCE CODE
All of the source code materials can be found in the UniHub submission link titled “Additional Files”. It has
been submitted in a zip folder and includes-

HTML Files
CSS Files
JavaScript Files
Image Files
Video Files
Audio Files
.eps Files

The documents are separated into folders. To run the website, all the JS, CSS, and HTML files need to
be removed from their individual folders and kept in the same folder, for it to execute and run properly.

APPENDIX 3 – SELENIUM IDE TESTING

TEST – 1

Page 103 of 119

The first test was to test the homepage. This included navigating through all the different sections of the
homepage.

The test was recorded and then tested, and no errors were found.

TEST – 2

The next test was to check if a user can successfully discover an artist on the discover page. The test
was successfully completed without any errors.

TEST – 3

Page 104 of 119

This was the most important test. It was to ensure that the user can go through the process of creating a
playlist.

This test included the process of opening the ‘Make a Playlist section, viewing the slideshow of
instructions, clicking “Create Playlist”, going through the process, saving the playlist, and viewing artist
recommendations.

The whole process went through smoothly with no errors in functionality.

APPENDIX 4 – ETHICAL SCREENING FORM

Page 105 of 119

Page 106 of 119
APPENDIX 5 – ETHICAL APPROVAL FORM

Page 107 of 119

Page 108 of 119
Page 109 of 119
APPENDIX 6 – RESEARCH PHASE CONSENT FORM

Page 110 of 119

APPENDIX 7 – USABILITY TESTING CONSENT FORM

Page 111 of 119

APPENDIX 8 – HUERESTICS EVALUATION CONSENT FORM

Page 112 of 119

APPENDIX 9 – MOODIFY USER SATISFACTION SURVEY

Page 113 of 119

Page 114 of 119
Page 115 of 119
APPENDIX 10 - MOODIFY USER SATISFACTION SURVEY RESPONSES

Page 116 of 119

Page 117 of 119
Page 118 of 119
Page 119 of 119