README.md

Music Recommendation System

A unsupervised learning model which analyses playlists and gives recommendations. Dataset used is FMA. Done by Sunil Boopalan from SRMIST.

Why

We created this project to see if we can actually understand the musical patterns of a listener with their playlist as source and what factors are really useful in determining the taste and interest of the listener.

Table of Contents

1. Next Steps
2. Installation
3. Run it

Next steps

If do not have jupyter and python visit Install Jupyter and Python

If you have them, proceed with the below steps.

1. Clone the repo

   $ git clone https://github.com/Sarathisme/music-recommendation-system.git

2. Visit the Run It section

Instal Jupyter and Python

1. Clone this repo to get the .ipynb files

   $ git clone https://github.com/Sarathisme/music-recommendation-system.git

2. Install python from https://www.python.org/downloads

3. If you already have jupyter in your machine, skip the next step.

4. Install jupyter either from conda or pip

   If you dont have conda installed, get it from https://docs.continuum.io/anaconda/install/

   From conda

   $ conda install -c conda-forge jupyterlab
   $ conda install -c conda-forge notebook

   From pip (pip is auto installed when you install python)

   $ pip install jupyterlab
   $ pip install notebook

Run it

Unfortunately at this point we do not have a .tar or a pickle file for you to quickly plug and play the code.

1. Go to the cloned folder

   $ cd path/to/code

2. Run setup.py to get the dataset installed and extracted into the project folder.

   $ python setup.py

3. Open jupyter notebook

   $ jupyter notebook

4. Open Music Recommendation System (Data Processing and Analysis).ipynb for data processing and analysis

5. Open Music Recommendation System (Machine Learning).ipynb for machine learning. This also has the recommendations code.