Abstract: Hit Song Science concerns the possibility of predicting whether a song will be a hit before its distribution using automated means such as machine learning software. This has motivated to dig deeper to unravel how different audio features would help to predict if a song would feature in the Billboard Top 100 Chart and build a two-way usability model - both for the musicians composing the music and the labels broadcasting it. The work in this paper also aligns with our team’s vision of exploring real-world applications of machine learning techniques and making them useful in common domains. In this paper, prediction models on data from Million Songs Dataset (MSD), Billboard, and Spotify using machine learning techniques have been explored, and low-level & high-level feature engineering techniques are applied. Finally, a comparison of their performances using various performance metrics has been carried out.