Predicting Song Popularity by Analyzing Audio Features of Spotify Bengali Tracks across Diverse Genres

At present, Spotify has become the most used music streaming platform, and a song that is popular song on Spotify is likely to be commercially successful. In the ever-evolving music industry, the ability to predict the potential popularity of a song before producing it can make impactful changes, thus leading to commercially successful music production. In this research work, we have introduced a methodology that predicts if a Bengali song is going to be popular on Spotify, utilizing a robust dataset derived from Spotify using the Spotify Web API. The dataset comprises various audio features of songs across multiple genres. We employed several machine learning models to forecast song popularity. Specifically Logistic Regression, Random Forest, K-nearest neighbors (KNN), Gaussian Naive Bayes, Decision Tree, XGBoost, and Support Vector Machine (SVM). Where XGBoost and SVC give promising results, we have also shown the features importance for XGBoost and SVC to highlight the important features to detect a song’s popularity. Our aim is to provide insights for more commercially successful music production in the Bengali music industry by analyzing data from the Spotify Web API.