Have you noticed that your preference is never constant? If you notice this happening whenever you listen to music, it turns out you are not alone and it happens a lot. According to a study conducted by Ole Adrian Heggli, Jan Stupatcher, and Peter Vuust, our music preferences change five (5) times within a 24 hour period. The 24 hour period is subdivided into the following:
- late night/early morning
The Research Study in a Nutshell
The study was conducted by analyzing audio features from over two (2) billion music streaming events from Spotify. The analysis was made possible by integrating an artificial neural network with Spotify’s API. The study did not collect personal data from individuals. The neural network only collected the metadata of the songs that were listened to, among the metadata collected are:
- beat strength
The research highlighted several interesting results, these are:
- The loudness of the music commonly played increases towards the end of the morning and stays steady and only dips at night.
- Tempo and Danceability increase during the night after a constant lull in the afternoon.
- In the morning, the music preference has an increased loudness, valence, and energy with a decreased tempo.
- In the afternoon, tempo increases while beat strength and danceability increase to average values.
- The tempo, beat strength, and danceability reaches their peak during the evening.
- At night, both tempo and loudness fall into their lowest value.
- During the late-night/early morning, the tempo and loudness are the same as that during the night, however, there is an increase in valence and energy.
Own playlist? or let the algorithm choose?
In another branch of the study, the researchers determined if the users are conscious of the type of music they play or they let the algorithm choose for them? With the 89 participants, the research showed that throughout the day with the exception of late night/early morning, people would prefer to play their own playlist rather than letting the Spotify algorithm choose. However, during the late-night or early morning, this awareness on choosing their own playlist over that of Spotify’s decreases.
The hallmark of this research is not only on the results but also on the process and how technology can predict our choices based on data points that are collected by these apps.