Predicting Song Performance on Music Platforms

Model trained on song characteristics to predict the binary classifcation of whether a song will do propritonally better on Spotify or Youtube

This project aims to segment certain types of music consumers into different music platforms.

It will do this by analyzing the question of whether a song will proportionally do better on Spotify or Youtube based on the traits associated with the song. Traits include both objective and subjective statistics such as duration of the song vs danceability of the song.

Technologies Used:

• Sci-kit Learn Machine Learning Models (logistic regression, KNN classifier, decision tree classifier, and unsupervised agglomerative clustering)
• Matplotlib and Seaborn Diagrams
• Jupyer Notebook to run Python and Format markdown
• Pandas and Numpy for data storage and manipulation

Dataset:

https://www.kaggle.com/datasets/salvatorerastelli/spotify-and-youtube

Dataset consists of songs from various artists in the world and for each song the following is present:

• Several statistics of the music version on Spotify, including the number of streams
• Several statistics of the official music video on Youtube, including the number of views

Viewing/Running Project:

• Individual Code blocks can be run in ipnyb file
• Comments and Analysis are layed out in markdown blocks throughout file