The evolution of the definition of “popular music” in the Internet era is directly attributable to the pervasive adoption of digital distribution methods for media and other types of information. This study uses machine learning to forecast a song’s commercial viability on Spotify. Various machine learning classification algorithms, including KNN, RF, and DT, will be trained and evaluated with an 80:20 split to forecast the popularity of the 232726 songs in the Kaggle database. Compared to other classification algorithms, such as the K Nearest Neighbors classifier (77% accuracy) and the Decision Tree classifier (70% accuracy), the results indicate that RF (92%) has the highest performance in predicting the appeal of music. The RF outperforms other methods in predicting the popularity of a composition within a given genre.