Predicting Music Popularity: A Machine Learning Approach Using Spotify Data

Shuo Jiang

2024

Abstract

In today's world, with the continuous advancement and application of streaming technologies, music has become ubiquitous and is increasingly integrated into the daily lives of people. This paper examines the application of machine learning algorithms for predicting music popularity through an extensive dataset sourced from Spotify, comprising 114,000 songs recorded over two decades. Traditional methods of predicting song success have often been subjective and inaccurate; however, advancements in artificial intelligence (AI) offer new avenues for improvement. This paper employed three machine learning models—Random Forest Regressor, Simple Linear Regression, and Gradient Boosting Machines—to analyze various audio features and their influence on song popularity. The Random Forest Regressor surfaced as the most effective model, capturing complex relationships within the data and achieving a respectable R² score. The findings highlight key predictors of popularity, including danceability, energy, and loudness, while also revealing challenges in accurately forecasting songs at both ends of the popularity spectrum. This research highlights the significance of incorporating various elements, including marketing tactics and social media engagement, in addition to audio characteristics, to improve predictive accuracy. Ultimately, the study showcases the capability of machine learning methods in grasping the intricacies of music popularity dynamics.

Paper Citation

in Harvard Style

Jiang S. (2024). Predicting Music Popularity: A Machine Learning Approach Using Spotify Data. In Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM; ISBN 978-989-758-738-2, SciTePress, pages 324-328. DOI: 10.5220/0013330000004558

in Bibtex Style

@conference{mlscm24,
author={Shuo Jiang},
title={Predicting Music Popularity: A Machine Learning Approach Using Spotify Data},
booktitle={Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM},
year={2024},
pages={324-328},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013330000004558},
isbn={978-989-758-738-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM
TI - Predicting Music Popularity: A Machine Learning Approach Using Spotify Data
SN - 978-989-758-738-2
AU - Jiang S.
PY - 2024
SP - 324
EP - 328
DO - 10.5220/0013330000004558
PB - SciTePress