In traditional music applications or websites, the number of music is so large that it is difficult for the average user to find the best music that will suit their current mood and needs. It may not be the case for people who have a relatively small catalog of music they listen to. However, since music is so large and the raw number of categories that exist today is steadily increasing, it would be more efficient to have a personalized list rather than manually go through such a vast array of music. It can prove even more difficult with a poorly designed music website lacking attractiveness. As a solution, I have designed and implemented a recommendation system for our music rating website, which will help increase the beauty of the website and predict the appropriate songs for the user based on key attributes and qualities of music that has been recently played.