🎶 Spotify Tracks Analysis

---

Is there any musical characteristic that can explain the popularity of a song in Spotify?

Introduction

Using the dataset stored in Kaggle (https://www.kaggle.com/datasets/yashdev01/spotify-tracks-dataset) I seek to answer if the popularity of a song on Spotify can be explained through its musical characteristics

This project explores a Spotify song dataset containing 125 different genres, to find relationships between the acoustic characteristics of the tracks and their popularity. The goal is to understand what attributes are associated with the most popular songs and what differentiates them from other songs.

Conclusions

Music does not follow a clear pattern in terms of popularity, it is evident that the most popular songs within this dataset are:

• Unholy (feat. Kim Petras) - "dance" genre
• Quevedo: Bzrp Music Sessions, Vol. 52 "hip-hop" genre
• I'm Good (Blue) - "dance" genre
• Bachata - "latin" genre
• Me Porto Bonito - "Latin" genre

Different genres, with a danceability above 0.56, a different mode, in some major, and in some minor, and a quite varied valence. It is to be expected that the characteristics of these songs are not exactly the same for each of them, but there is no clear pattern either. In fact, None of the 5 most popular songs are within the 5 most popular genres, these are:

• pop_film
• k-pop
• chill
• sad
• grunge

This means that the most popular genres have many songs that are listened to a lot but none become extremely popular, on the contrary, the most popular songs are extraordinary cases of genres that do not usually stand out, but they have one or two hits so big that they end up becoming the most popular song genres.

When studying the correlation between the characteristics, there was no evidence that any of these musical characteristics influenced the popularity of a song.

Even though characteristics (not including popularity) relate to and affect each other, popularity never seems affected by tempo, valence, danceability, or even mood of the song, these musical characteristics take a backseat in today's music industry. Marketing seems to have much more weight, the virality, the controversy and the weight generated by conversations through social networks than the musical structure of the song itself.

It is clear that it seems an almost impossible task to predict through this data whether a song may or may not be popular; it is a natural human behavior to be influenced not only by musical richness when choosing a song but also by the environment in which we find ourselves.

---

💻 How to use it?

1. Create the virtual environment in your terminal with python3 m-venv venv
2. Enter the virtual environment with source venv/bin/activate To exit the environment simply type deactivate
3. Install the requirements by reading the requirements archive with pip install -r requirements.txt
4. Run the file using python\_3 spotify_tracks_analysis.py