Million Song Dataset

Introduction

This is a big data project which have been done by utilising the Million song dataset from the following link.

Aim

The objective of the project was song recommendation and year prediction by using metdata of the songs from the dataset
The project involves following parts

1) Bringing Data into proper format->make_csv.R
2) Data Cleansing
3) Year Prediction -> code in years_predict folder
4) Song Recommendation-> code in song_recommendation folder
5) Running the algorithm on Hadoop using Big Data Service Qubole
6) Data Visualisation -> visualisation folder

Note-
1)In Data cleansing NA's and NAN's were removed and the values in column whose data was missing it was replaced by mean of the column like in  song.hotttness column some values were zero
2) you can have a look at song_dataset.csv to look at the metadata of the songs which was used for the project

Teammates

Avik Basu
Taranveer Singh
Sahil Govel
Raj Biswas
Akash Lodha
Shikhar Sharma
Sresht Agrawal