Spotify Ecosystem
What artists are the most valuable collaboration within the spotify eco system? This is an important question that could be asked by many industry record label teams to identify best performing collaborations to expose artists to possible new audiences to continue to grow. Analyzing the artist collaboration network will give a better understanding of the roles that collaborations structures play in introducing new audiences and popularity. The answer to this question will produce the optimal partnerships amongst musicians to achieve the greatest outreach while still having an intended audience.
The dataset that was used for this project was provided via kaggle(https://www.kaggle.com/datasets/maharshipandya/-spotify-tracks-dataset). This dataset provides several relevant elements such as track_id , artists, track_name, album_name, popularity, track_genre, and danceability. The main element we will analyze is the artist column to further understand the network.
In the network created each node is used to represent an individual musical artist which appeared in the spotify dataset. The dataset contains several artists and uses the combined songs of an artist to determine the relationships between artists.Each artist receives one individual link per track shared amongst artists. These are for edges which are weighted by the amount of times an artist has collaborated with one another.
In our analysis importance was constructed using several methods to determine the influence of an artist. J Balvin is one of the most influential within the dataset when using the page the weighted degree formula which tell us the total number of collaborations an artist has but just that doesn’t define the importance of an artist, so we used other methods to also determine importance such as betweenness centrality which identified people like anitta with a high score of 0.018819 identify them as a bridge between artist. Lastly we also looked at page rank which identified the value of artists who are most influential within their own networks. In this case the biggest stand out was Wolfgang Amadeus Mozart. In our analysis we will look at all of these factors and determine our own factor of importance with all these different measurements being a matrix of importance.
Get JuniorBenitez’s stories in your inbox
Join Medium for free to get updates from this writer.
Within all the metrics, J Balvin ranks very highly using the final custom-created importance metric, which takes 25% from each of the other metric categories to create an overall importance score. In this case, J Balvin ranks the highest, making him the most desirable artist to collaborate with within this dataset from Kaggle.
This is a network diagram of the 100 most important people. This reveals 3 main clusters that describe 3 communities based on the artists seen. The middle left cluster focuses on EDM/pop music and is at the center of connecting all the communities. Within this cluster, three main bridge artists that also have large popularity are Illenium, Marshmello, and J Balvin. This middle bottom cluster focuses on Indian and Bollywood music. Pritam stands out the most in this cluster with the most popularity and the highest importance, with Pritam also serving as a bridge between multiple artists. The last far right light blue cluster has a focus on classical music, with artists such as Mozart being the most popular and the central focus of the cluster. Andrea Bocelli is an outlier within this cluster that represents a bridge between pop and classical music artists.
The top 10 bridge artists highlight the artists that connect other artists together by using the betweenness centrality metric. It captures the amount of collaborations an artist has between other artists and highlights artists like Ferrugem and Anitta that rank high, showing their critical role in linking separate groups together. Andrea Bocelli was also noted within the network graph and clearly displays himself halfway between two clusters of classical and pop/EDM music.
The collaborations an artist has were measured by the weighted degree metric, which represents the total number of collaborations an artist has. Artists like J Balvin and Mozart rank highest here, indicating a large volume of collaborations with relevant artists. This suggests the importance within a community that an artist has, such as J Balvin being a central focal point within the EDM and pop genre cluster. These artists don’t necessarily dominate in connecting music genres, but they show the power they hold within their own genre. Some of these characters are also bridge artists, showing influence in and out of their genre by connecting communities, making them more influential, like J Balvin.
The importance rating is a custom score created by me that takes into account 25% of each metric to give artists a final ranking that determines who the most influential artists are overall across all categories. Artists like Pritam and Mozart stand out the most here, consistently ranking near the top in every category. While other artists like J Balvin and ILLENIUM dominate in one or two categories, they still appear slightly lower due to weaker performance in others. Overall, this demonstrates that Pritam, Mozart, and J Balvin are among the most influential artists, each showing different levels and types of influence within the network.
AI tools were used to assist in the coding process for this project. To ensure the results were reasonable, I carefully reviewed the dataset and manually checked portions of the output by performing calculations in Excel to verify that the code produced accurate values. During this validation process, some potential bias within the dataset was identified, which will be discussed in the following section.
One limitation of this analysis is the scope of the dataset used. The Kaggle dataset provides a good starting point, but it only contains a sample of songs rather than the full Spotify catalog. Because of this, some artists and collaborations may be missing, which can skew the results. Additionally, collaborations are defined only by artists appearing on the same track, which may not always represent a direct collaboration. The analysis could be improved by collecting data using the Spotify API, but gathering and processing that amount of data would require more time and computing resources than were available.
GitHub: https://github.com/Junior-JB/INST-414/blob/main/module%202.py
