Using Spotify’s Web API to Examine What Audio Features Impact a Song’s Popularity
Abstract
Spotify is one of the most popular music streaming services in the world and has over one hundred million different music tracks available to its users. Being able to predict what audio features contribute to a song's popularity could be useful for musical artists and record labels as well as music listeners. Spotify offers a web application programming interface (API) that allows users to pull data on a variety of audio features for individual songs. The goal of my capstone project was to acquire and clean a data set using this API in python. Regression models were fit to predict a song's popularity using the audio features of the song such as the song's instrumentalness, energy, and key. None of the models created were very accurate; the highest R-squared value obtained was around 0.1. However, some of the audio features such as the key the song is in and the danceability of the song seemed to have a greater impact on a song's popularity than other features. These results indicate that some audio features have a greater impact than others but many other factors beside the audio features most likely have a large impact on a song's popularity such as the lyricism or artist.
College
College of Science & Engineering
Department
Mathematics & Statistics
Campus
Winona
First Advisor/Mentor
Tisha Hooks
Start Date
4-19-2023 9:00 AM
End Date
4-19-2023 10:00 AM
Presentation Type
Poster Session
Format of Presentation or Performance
In-Person
Session
1a=9am-10am
Poster Number
33
Using Spotify’s Web API to Examine What Audio Features Impact a Song’s Popularity
Spotify is one of the most popular music streaming services in the world and has over one hundred million different music tracks available to its users. Being able to predict what audio features contribute to a song's popularity could be useful for musical artists and record labels as well as music listeners. Spotify offers a web application programming interface (API) that allows users to pull data on a variety of audio features for individual songs. The goal of my capstone project was to acquire and clean a data set using this API in python. Regression models were fit to predict a song's popularity using the audio features of the song such as the song's instrumentalness, energy, and key. None of the models created were very accurate; the highest R-squared value obtained was around 0.1. However, some of the audio features such as the key the song is in and the danceability of the song seemed to have a greater impact on a song's popularity than other features. These results indicate that some audio features have a greater impact than others but many other factors beside the audio features most likely have a large impact on a song's popularity such as the lyricism or artist.
