Music Recommendation Based on Emotion

Description

As part of the 2024 SUTD Computational Data Science Module, we decided to explore Music Emotion Recognition and Retrieval to better understand how audio processing works and how we could develop models that can recommend songs that are emotionally similar.

Report

Please click on this link to view our report that documents our end-to-end process in this project.

Table of Contents

• Installation
• Usage
• Contributing

Installation

The following project requires OpenSMILE (required for dashboard directory) and Essentia to run the project.

1. Refer to the OpenSMILE Documentation to compile the OpenSMILE package from source.
2. Refer to the Essentia Documentation to install the Essentia library.
3. Git clone this repository to the desktop
4. Set up a Python Virtual Environment inside the root directory of the GitHub Repository
5. Activate the virtual environment and install the package dependencies using the following command pip install -r requirements.txt

Usage

The project is composed of 2 main directories - dashboard and research The dashboard directory consists of files to run the Streamlit Web Application Demo while the research directory consists of files, mainly notebooks, in which we conducted our experimentation and analysis.

Running the Streamlit Web Application

• Run the Streamlit demo via the following command streamlit run ./dashboard/UI/app.py

Contributions

The following project was completed by

• Lim Fuo En
• Anthony Lim
• Issac Jose Ignatius
• Koh Jia Jun
• Timothy Wee