PROJECT REPORT

ON
SPOTIFY DATA ANALYSIS

This report serves as documentation for the R PROGRAMMING course's lab work,
focusing on the exploration of

BTECH IN COMPUTER SCIENCE ENGINEERING.

Prepared by
Vamshi Reddy 22911A05N2
Y Abhishek 23911A05R1
Roshan Vishwakarma 23915A0518
Himanshu Sai 23915A0519

Under the guidance of

Mrs. BABITHA madam

VIDYA JYOTHI INSTITUTE OF TECHNOLOGY

Department Of Computer Science and Engineering
(Aziznagar Gate, Chilkur Road, Hyderabad- 500075, Telangana, India.)
 CONTENTS

S.NO DESCRIPTION PAGE NO

1 Abstract 1

2 Introduction 2

3 System Requirements 3

4 Introduction to R language Data Analysis 4

5 R librariess 7

6 Coding 8

7 Screenshots 14

8 Conclusion 17
 ABSTRACT

This documentation presents a detailed overview of a

comprehensive analysis project focused on Spotify data. The
primary objective of this project is to delve into the intricate
musical patterns and user preferences within the vast and
diverse Spotify ecosystem. Leveraging the richness of Spotify's
dataset, encompassing user listening habits, track attributes,
and various other metadata, our analysis aims to uncover
meaningful insights into the world of music consumption.
The project encompasses several key components, including
data collection, preprocessing, and exploration, utilizing
cutting-edge data science and machine learning techniques.
Our analysis delves into understanding the dynamics of user
behaviour, examining factors influencing song popularity, and
identifying patterns that contribute to personalized music
recommendations.

1|Page
 INTRODUCTION
In an era characterized by an abundance of digital music streaming
platforms, Spotify stands as a frontrunner, captivating millions of users
with its vast catalogue and personalized recommendations. This
project embarks on a journey to unravel the intricate tapestry of
musical preferences and behaviour’s embedded within Spotify's
extensive dataset. By employing advanced data analysis and machine
learning techniques, we aim to discern patterns, trends, and insights
that illuminate the diverse ways in which users interact with music on
this platform.
We delve into the realm of machine learning, employing sophisticated
algorithms to uncover hidden patterns and provide insights into the
underlying structures of Spotify's vast musical landscape. By doing so,
we not only aim to contribute to the growing field of music data
analytics but also to enhance our understanding of the intricate
interplay between technology, data, and the human experience of
music consumption.

2|Page
 SYSTEM REQUIREMENTS
HARDWARE REQUIREMENTS:
Operating System: windows 10

Processor (CPU): Intel(R) core(TM) i3 7020U CPU

Memory (RAM): 8 GB

SOFTWARE REQUIREMENTS:
Integrated Development Environment (IDE) or Text Editor: R studio

Programming Languages: R language

3|Page
 INTRODUCTION TO R LANGUAGE
&
DATA ANALYSIS

R, a powerful and versatile programming language, has emerged as a

cornerstone in the field of data analysis and statistics. Designed with a focus on
statistical computing and data visualization, R provides a robust platform for
professionals and researchers to explore, analyses, and interpret complex
datasets. This introduction aims to provide an overview of R's significance in data
analysis, emphasizing its strengths, key features, and applications in unlocking
insights from diverse datasets.
R's open-source nature and extensive package ecosystem make it a
preferred choice for statisticians, data scientists, and analysts. Its flexibility and
scalability empower users to handle a wide range of data types and sizes, from
small-scale exploratory analyses to large-scale, complex modelling tasks. The
language's syntax is intuitive, enabling both beginners and experienced users to
quickly grasp its fundamentals and begin their journey into the world of data
analysis.
One of R's defining features is its emphasis on visualization. With
dedicated libraries such as ggplot2, users can create compelling and informative
graphs and charts that facilitate a deeper understanding of data patterns and
trends. R's visualization capabilities extend beyond static plots, with interactive
visualizations becoming increasingly accessible through packages like Shiny.
In the realm of statistical analysis, R offers a comprehensive suite of tools
for hypothesis testing, regression analysis, and machine learning. Its statistical
functions are continually updated and refined by a vibrant community of
contributors, ensuring that R remains at the forefront of statistical
methodologies and techniques.

4|Page
 Furthermore, R seamlessly integrates with other data analysis tools and
platforms, making it an asset in diverse data ecosystems. Whether used
independently or in conjunction with tools like RStudio or Jupiter notebooks, R
provides a flexible environment for coding, analysis, and reporting.
This introduction sets the stage for a deeper exploration of R's capabilities in data
analysis.
As we navigate through the various aspects of R programming, statistical
modelling, and data visualization, we aim to showcase the language's prowess
in transforming raw data into meaningful insights, empowering users to make
informed decisions and discoveries in the dynamic landscape of data analysis.

5|Page
 STEPS IN DATA ANALYSIS
Data analysis involves a systematic process of inspecting, cleaning,
transforming, and modelling data to extract meaningful insights. Here
are the typical steps in a data analysis workflow:
Define the Problem or Research Question:
Clearly articulate the problem or question you aim to address through
data analysis. This sets the context for the entire process.
Data Collection:
Gather relevant data from various sources, ensuring it aligns with the
objectives of your analysis. This could involve surveys, experiments,
sensors, databases, or other data repositories.
Data Cleaning:
Examine the dataset for errors, missing values, and inconsistencies.
Clean and preprocess the data to ensure its quality and reliability.
Exploratory Data Analysis (EDA):
Conduct an initial exploration of the data to understand its structure
and characteristics. This involves summary statistics, data
visualizations, and identifying patterns or outliers.
Formulate Hypotheses:
Based on your initial exploration, develop hypotheses or questions
that you aim to answer through statistical testing or further analysis.
Statistical Analysis:
Apply appropriate statistical methods to test hypotheses, identify
relationships between variables, and derive meaningful insights. This
may involve descriptive statistics, inferential statistics, or machine
learning algorithms.

6|Page
R LIBRARIES
In a Spotify data analysis project using R, various libraries can be
leveraged to efficiently handle, analyze, and visualize the data. Here
are some commonly used R libraries that might be applicable in such
a project:
• tidyverse:
The tidyverse is a collection of R packages, including dplyr,
ggplot2, tidyr, and others. These packages provide a cohesive
and consistent set of functions for data manipulation, cleaning,
and visualization.
• ggplot2
ggplot2 is a powerful and widely-used R package for creating
static and dynamic data visualizations. Developed by Hadley
Wickham, ggplot2 is based on the "Grammar of Graphics"
framework, which provides a structured and consistent
approach to creating plots. This package is particularly popular
for its simplicity, flexibility, and ability to produce high-quality
graphics.
• dplyr
dplyr is a widely used R package that provides a set of functions
for data manipulation and transformation. Developed by Hadley
Wickham, dplyr is part of the tidyverse collection and is designed
to work seamlessly with data frames, offering a consistent and
intuitive syntax. It simplifies and streamlines common data
manipulation tasks, making it a go-to tool for analysts and data
scientists.
• tidyr
tidyr is another essential R package that is part of the tidyverse
collection, alongside dplyr. While dplyr focuses on data
manipulation, tidyr is designed for data tidying and reshaping. It
provides a set of functions for restructuring data frames to make
them more suitable for analysis and visualization. Developed by
7|Page
 CODING
install.packages("tidyverse")

# Load required libraries

library(tidyverse)

spotify_data <- read.csv("C:\\Users\\Lenovo\\Desktop\\r-project\\spotify_data.csv")

# Display the structure and first few rows of the dataset

str(spotify_data)

head(spotify_data)

# Check for missing values

summary(spotify_data)

# Load necessary libraries

library(dplyr)

# Assuming spotify_data has been properly loaded and defined previously

# Load necessary libraries

library(ggplot2)

# Count occurrences of each genre

genre_counts <- spotify_data %>%

group_by(genre) %>%

summarise(count = n()) %>%

arrange(desc(count))

# Create a bar chart for Genre Counts

ggplot(genre_counts, aes(x = reorder(genre, -count), y = count, fill = genre)) +

geom_bar(stat = "identity") +

labs(title = "Genre Counts on Spotify", x = "Genre", y = "Count") +

theme(axis.text.x = element_text(angle = 45, hjust = 1)) +

scale_fill_brewer(palette = "Set3")

# Load necessary libraries

library(ggplot2)

# Assuming your data is loaded and named as 'your_data'

# Create a histogram for 'bpm'

8|Page
ggplot(spotify_data, aes(x = bpm)) +

geom_histogram(binwidth = 5, fill = "skyblue", color = "black") +

labs(title = "Distribution of Beats Per Minute (bpm)", x = "BPM") +

theme_minimal()

# Count occurrences of each artist

top_artists <- spotify_data %>%

group_by(artist) %>%

summarise(count = n()) %>%

arrange(desc(count)) %>%

top_n(10, count)

# Create a bar chart for Top 10 Artists

ggplot(top_artists, aes(x = reorder(artist, -count), y = count)) +

geom_bar(stat = "identity", fill = "green", color = "black") +

labs(title = "Top 10 Most Frequent Artists", x = "Artist", y = "Count") +

theme(axis.text.x = element_text(angle = 45, hjust = 1))

# Count occurrences of each song title and filter for Top 5 Songs

top_songs <- spotify_data %>%

group_by(title) %>%

summarise(count = n()) %>%

arrange(desc(count)) %>%

slice(1:5) # Filter for Top 5 Songs

# Create a bar chart for Top 5 Songs

ggplot(top_songs, aes(x = reorder(title, -count), y = count)) +

geom_bar(stat = "identity", fill = "green", color = "black") +

labs(title = "Top 5 Most Frequent Songs", x = "Song Title", y = "Count") +

theme(axis.text.x = element_text(angle = 50, hjust = 1))

# Display top 10 artists

print(top_artists)

# Display top 10 songs

print(top_songs)

9|Page
# Assuming 'your_data' is your dataset

summary_stats <- spotify_data %>%

summarise(

mean_bpm = mean(bpm),

median_bpm = median(bpm),

min_bpm = min(bpm),

max_bpm = max(bpm),

mean_nrgy = mean(nrgy),

median_nrgy = median(nrgy),

min_nrgy = min(nrgy),

max_nrgy = max(nrgy),

mean_dnce = mean(dnce),

median_dnce = median(dnce),

min_dnce = min(dnce),

max_dnce = max(dnce),

mean_dB = mean(dB),

median_dB = median(dB),

min_dB = min(dB),

max_dB = max(dB),

mean_live = mean(live),

median_live = median(live),

min_live = min(live),

max_live = max(live),

mean_val = mean(val),

median_val = median(val),

min_val = min(val),

max_val = max(val),

mean_dur = mean(dur),

median_dur = median(dur),

min_dur = min(dur),

max_dur = max(dur),

10 | P a g e
 mean_acous = mean(acous),

median_acous = median(acous),

min_acous = min(acous),

max_acous = max(acous),

mean_spch = mean(spch),

median_spch = median(spch),

min_spch = min(spch),

max_spch = max(spch),

mean_pop = mean(pop),

median_pop = median(pop),

min_pop = min(pop),

max_pop = max(pop)

# Display the calculated summary statistics

print(summary_stats)

library(dplyr)

library(ggplot2)

# Calculate summary statistics for numerical columns

summary_stats <- spotify_data %>%

summarise(

mean_bpm = mean(bpm),

median_bpm = median(bpm),

min_bpm = min(bpm),

max_bpm = max(bpm),

mean_nrgy = mean(nrgy),

median_nrgy = median(nrgy),

min_nrgy = min(nrgy),

max_nrgy = max(nrgy),

# Add other columns in a similar manner for their summary statistics

11 | P a g e
# Convert summary statistics to a tidy format for plotting

summary_stats_tidy <- summary_stats %>%

gather(statistic, value)

# Plotting summary statistics using bar chart

ggplot(summary_stats_tidy, aes(x = statistic, y = value, fill = statistic)) +

geom_bar(stat = "identity", position = "dodge") +

labs(title = "Summary Statistics of Numerical Columns", y = "Value") +

theme_minimal()

# Assuming 'your_data' is your dataset

specific_year_data <- spotify_data %>%

filter(year == 2019) # Change the year to the specific year you're interested in

# Count occurrences of each artist in the specific year

popular_artists_2019 <- specific_year_data %>%

group_by(artist) %>%

summarise(count = n()) %>%

arrange(desc(count)) %>%

top_n(5, count) # Get the top 10 artists for 2020

library(ggplot2)

# Assuming 'popular_artists_2020' contains the top artists for the year 2020

# Create a bar chart for Top 5 Artists in 2020

ggplot(popular_artists_2019, aes(x = reorder(artist, -count), y = count)) +

geom_bar(stat = "identity", fill = "magenta", color = "black") +

labs(title = "Top 5 Artists in 2019", x = "Artist", y = "Count") +

theme(axis.text.x = element_text(angle = 45, hjust = 1))

library(ggplot2)

library(dplyr)

library(tidyr)

# Assuming 'your_data' is your dataset

12 | P a g e
# Gather musical attribute columns into key-value pairs

your_data_long <- spotify_data %>%

gather(key = "Attribute", value = "Value", -c(title, artist, genre, year))

# Create a line plot for musical attributes over years

ggplot(your_data_long, aes(x = year, y = Value, color = Attribute, group = Attribute)) +

geom_line() +

labs(title = "Trends of Musical Attributes Over Years",

x = "Year", y = "Value",

color = "Attribute") +

theme_minimal()

13 | P a g e
 SCREENSHOTS
Genre Count in the dataset

Top Artists from 2010-2019

14 | P a g e
Top songs

Top artist

15 | P a g e
Top artist in 2019

Trends Of Musical Attributes Over Years

16 | P a g e
Conclusion:
The Spotify data analysis project conducted using the R programming language
has provided valuable insights into the intricate world of music consumption and
user preferences within the Spotify platform. Through a systematic and
comprehensive approach, we navigated the vast dataset, employing a
combination of data manipulation, statistical analysis, and visualization
techniques to uncover patterns and trends.
The initial stages of the project involved data collection from the Spotify Web
API using the spotifyr package, allowing us to access detailed information about
tracks, albums, and user interactions. The data was meticulously cleaned and
preprocessed using the dplyr and tidyr packages, ensuring its quality and
relevance for subsequent analysis.
Exploratory data analysis (EDA) unveiled interesting patterns in user behavior,
highlighting trends in song popularity, genre preferences, and temporal
variations. Visualizations created with the ggplot2 package provided a clear and
intuitive representation of these findings, allowing for a deeper understanding
of the data.
Statistical analyses, facilitated by dplyr and other relevant packages, further
elucidated relationships between variables, enabling the identification of factors
influencing the popularity of certain tracks. Machine learning algorithms
implemented with the caret package contributed to predictive modeling,
enhancing our ability to forecast user preferences and trends.
The project also delved into sentiment analysis using tm and other text mining
tools, providing insights into user sentiments expressed in comments or reviews.
This added a qualitative dimension to our understanding of user interactions
with music on Spotify.
As we conclude this project, it is evident that R, with its rich ecosystem of
packages and versatile capabilities, has proven to be an invaluable tool for
exploring and analysing Spotify data. The project's methodologies,
visualizations, and findings not only contribute to the broader field of data
analytics but also offer practical applications for improving the user experience
and shaping the future of personalized music recommendations within the
Spotify platform.

17 | P a g e