Employing Spotify to Track Global Moods

Date that article was published: February 20, 2021

Article Summary

Spotify users turn to music on a daily basis to lift their spirits (mood incongruency) or to align with their sentiments at that current moment (mood congruency). Pairing data science and Spotify app data reveals how global mood trends align with the characteristics of the app’s top 200 streamed songs. Employing a weighted calculation of the valence (musicality and mood of each song, including tempo and melodic features) and the stream frequency allowed for an estimated “mood of the week” metric that was used for analysis. Results display that in the first week of quarantine, global moods and positive song valence were high, but quickly dropped as Americans adjusted to the new normal. During winter holidays in both 2019 and 2020, positive song valence was also relatively high. Employing data science to answer questions about the “mood” through definable valence metrics can reveal more insights into consumerism and inform market strategies at large.

Future Applications

Drawing a correlation between overall global mood and listener song preferences demonstrates that music has the ability to reflect sentiments across culture and place. This unique span can be used in the realm of consumerism, where music could be paired with the marketing of products to bolster sales. In the music industry specifically, the weighted metric employed in the study about “global mood” could be used to identify the most favorable time periods to release an album or song. Such mood metrics and their corresponding soundtracks could even be used to increase productivity in the workplace.

Author Information

Will Mcconnell is a student studying History and Math. Outside of school he is interested in the areas of politics and philosophy. Other Medium articles he has written involved following the Georgia runoffs and voting. His diverse academic background confirms that data science can be learned by anyone, and displays its applicability and functionality in a wide array of disciplines!

Coding Technique Takeaways

commands_employed <- c("read_csv()","data.frame()","na.omit()","return()")
commands_employed

One interesting technique I noted was the use of the ‘$’ symbol to select a column from a data frame, which is syntactically different in R as compared to the use of the ‘.’ or ‘[]’ notation employed in Python. I also found the data.frame() function to make a data frame of values helpful, as well as the read_csv() function we have learned about in class for reading in files.

Iris Datatable with DT

Iris Plot

Related Links

Other articles about the intersection of music, consumerism, and metrics of happiness

A few links to related articles can be found below: 1.https://www.researchgate.net/publication/232593421_The_Influence_of_In-Store_Music_on_Wine_Selections
2.https://www.psychologytoday.com/us/blog/the-happiness-doctor/201705/measuring-happiness-how-can-we-measure-it
3.https://www.sciencedirect.com/science/article/abs/pii/S0022435917300520?via%3Dihub

Useful Data Science Practices Gleaned

The most helpful part of the article was how the author went about understanding a metric of success. We discussed in class how once an objective for a data science project is defined, a metric for success should be identified. In this specific case, the valence and musicality metric for each of the songs was weighted on the stream quantity for each song, and the weighted average became the metric to gauge global moods. The author clearly delineated this metric throughout the article, and came back to it to underscore how it was helping him to reach his overall goal.