Introduction

Spotify allows people to download their data. They have an incrdibly interesting way at analyzing songs, using AI to determine things that are not tangible, such as danceability, instrumentalness, etc. Here’s a description of each of the variables in this data set. The data can be downloaded here; https://www.kaggle.com/datasets/geomack/spotifyclassification?resource=download

Danceability. A measure of how well a song is made for dancing, rated on a scale from 0-1.

Energy. A perceptual measure of activity and intensity, typically fast, noisy, loud, etc. Scale from 0-1.

Key. The key of the song. 0 is C natural, 1 is C#, 2 is D, etc.

Loudness. The measure, in decibels, of how loud a song is.

Mode. Depicts whether the song is major or minor. Major is 1, minor is 0.

Speechiness. Measure of how much spoken word is in a song. Scale of 0-1.

Acousticness. A measure of how acoustic a song is, from 0-1.

Instrumentalness. Confidence measure of how instrumental a song is, scale of 0-1.

Liveness. Measure of if the song has an audience. Confidence measure scaling from 0-1.

Valence. Happiness meter. 0 is sad, angry, while 1 is happy, cheerful.

Tempo. Describes in BPM the tempo of the song.

Duration. Describes in ms the length of a song.

Time signature. Depicts the time signature of the song. Ranges from 3-7, indicating the time signatures of 3/4 to 7/4. eg., a 3 would indicate a 3/4 song.

Liked. Dummy variable on whether or not the user ‘liked’ the song, using spotify’s like feature.

What key has the worst danceability?

In this graph, we look at key vs. danceability. It would appear that the key of G# has the worst danceability. Weirdly enough, the key of Eb has the highest average danceability. We would assume that C, being the key of most pop songs, would have the highest, however many rock songs and metal songs use the key of Eb.

Do energy and tempo have a correlation?

In this graph we observe the natural pattern of tempo and energy. This makes sense, as a higher tempo would result in a more energetic feeling.

Comparing valence to tempo based on mode

Here we can see a very interesting chart. The major key, expectadley, has a line of best fit higher on the valence axis than minor. However, at a slower tempo, it would appear that the minor key has a higher valence.

If we think about this, minor keys do not always mean sad. The minor key is the key that most rock bands prefer. The simplicity of the Am key makes it easy to make a song over. With a major key, a slower tempo could mean that a instrumental player, such as a guitar player, would be playing with more ‘feel’. Bends, vibrato, sustained notes; things typically related with emotion. Very interesting indeed!

You can tell a lot about a person by what songs are in their playlist

This particular person has a pretty even split, but prefers songs with a major, or happier, key.

Instrumentalness and danceability compared by energy.

This individual’s preference of instrumental songs is generally not meant for dance. We can assume that these songs are classical music pieces, or songs that are slower and more defined. If it were danceable, we would assume that this person listens to EDM, dubstep, etc.