It's not exactly a thunderous revelation that in recent times we've all been practically run over by a name resonating like a musical tsunami: Taylor Swift. And I'll be honest, my obsession with her lyrical ability is so obvious that it would be easier to hide it in a packed stadium from The Eras than to search for all the hidden Easter eggs in "Look What You Made Me Do." It's as if she has a song ready for every micro-moment in life, whether you're in the dumps or at the peak of a wild party.
So, I got to thinking... could we create a kind of electronic guru, an artificial intelligence, that selects the perfect Taylor Swift soundtrack according to your current mood? And thus was born Dorothea AI, my modern version of a musical advisor, but with a healthy dose of technological cuteness. To train this beauty, I fed her with 233 songs, leaving aside the original versions and focusing solely on the editions stamped with the "taylor’s versions" seal (after all, we don't listen to stolen versions).
All of this wouldn't have been possible without the magic of TensorFlow and the hands-on approach of Python, which were the secret ingredients of the recipe. Now, if you want to take a peek at the result of this slightly crazy experiment, just click here.
Oh, and of course, if you're the curious type who likes to tinker with gears, the entire process is neatly stored in this digital vault. Who knows, maybe you'll be inspired to create the next AI that chooses songs based on the weather? The sky's the limit, my friend!
Top comments (0)