It will keep you updated about new stories and give you the chance to text me to receive all the corrections or doubts you may have. Follow me on Linkedin, where I publish all my stories B. If you liked the article and you want to know more about Machine Learning, or you just want to ask me something you can:Ī. If it is totally wrong, at least I had A LOT OF FUN working on it, which is the most important thing to me right now :D The assumptions seem reasonable to me, so I think that if it is not totally correct I may have got some general idea about how it worksĢ. I don’t know if the reverse engineering process is right because again, I don’t work for Spotify.ġ. High values of energy and negativity (or low values of valence) are responsible for the “ angst” mood. High values of danceability are responsible for the “ chill” mood that I have as a top oneĢ. The Positivity feature is low ( 46%), while the negativity is (of course) still above 50% ( 53%).Īs our assumptions is that Energy, Danceability and Valence are the three main features used in this case by Spotify we can assume that:ġ. The Danceability and the Energy features are way above 50% on average ( 62% and 63%)ī. Here’s a list of the things we can state:Ī. Here is my Palette:Īnd this is the moment of truth. In fact, if we look at the description of the “moods” of the first chapter, and we compare it with the features above the we can see that it is not so difficult to match the colors with features like energy, danceability and valence.Īnother good hint that points in that direction is the so called Spotify Palette, that adapts the features with the moods and give you a palette of colours, given the danceability, energy and valence of your Spotify. It is thus reasonable to assume that, as these features are so complex and informative, Spotify use them in their Wrapped Experience as well.

When we listen to a song we are able to determine the danceability, energy or acousticness (if we studied a little bit of music) but it is way more difficult to build a smart machine that is able to do this for us.

It is trivial to say that this is one of the greatest work of art of Spotify. Valence is a measure of musical positivenessĪll these metrics are value between 0 and 1.Energy is a measure of intensity and activity.Danceability indicates how much a song is danceable.Acousticness is a confidence measure of wether the track is acoustic.While this other article explains the features in a very detailed manner and use them to play with a recommender system.

Stats for spotify not accurate