You may say I'm the next hit artist, but I'm not the only one ?

Have you wondered how a karaoke machine decides your singing ability? As a singer, I thought that it was just a computer matching the notes I sing with the artist's notes. Turns out, data science has found itself yet another application, karaoke machine scoring systems. And the same application can be applied in the entire music industry, in any place where computers are being used. Picture yourself singing into a karaoke mic, watching the sound waves that generate on the screen in front of you. Your voice is taken in in the form of a digital signal, and usually, the computer measures your volume, pitch, and timing. This is the basic algorithm that karaoke systems are trained to use. Notice that only one part of this is your ability to match the notes of the original song. Some of the fancier machines even match your performance with the artist, as I had thought. So, how do you trick the system into thinking you're a good singer? Is it possible at all? The algorithm is trained in the following way. If you sing loudly, the algorithm automatically sees you positively. If the timing is correct, especially in the beginning, the algorithm favors you. You can observe these aspects in the wave frequencies that the screen shows. Third, knowing the exact words goes a long way. So, you don't really need to be the best singer or even trained to achieve a high score with these algorithms. You just need to use the correct dynamics and be confident! Data science also dominates the music industry in ways more than just personalized playlists. Have you noticed that a lot of similar songs are produced? That some of the hit songs are actually quite similar? The idea is that you are more comfortable with things that you are familiar with, so the chance of you buying a song that's similar is higher too. I always believed that there were just limited tunes and words, but it turns out that this is the trick that producers and artists use. They map out the loved tunes, genres, words, phrases, beats, and instruments, and apply those. There's usually a large research team behind every music production house, that uses traditional data science methods of logistic regression, random forest, and support vector machines. All fancy methods to understand people and capitalize on their preferences! So pull up your socks and nail that proposal song, otherwise, you might be #AbsolutelySingle!