A peep at the current state of jazz music through data from Spotify (part 1)
Descriptive analysis using data returned from genre-specific search
Abstract
Jazz has been thriving for over a century since its inception. After its heyday in 1920s to 1930s, it slowly moved away from mainstream. Since then, jazz artists have been incorporating elements from other genres, e.g., blues, R&B, funk and rock, and achieved various degree of attention and commercial success. However, jazz’s popularity never returns to anywhere close to what it used to be. This study attempted to describe the current state of jazz with data obtain through Spotify Web API and various other sources. Data quality, availability, characteristics, popularity of jazz tracks and followers of jazz artists were examined closely and discussed in detail. (Don’t miss the jazz playlist I created for this article!)
Motive
Jazz is my favourite genre. I was an amateur guitarist who used to perform jazz occasionally back in Hong Kong. I happen to know a bit of data analytics too. So why not do a project that combines my passion for both? Jazz is also tremendously significant, both culturally and musically. It is the “America’s classical music”. It is the greatest musical offering from America to the world. I believe it deserves more attention and appreciation than it has now.
Methodology
This study attempted to describe the current state of jazz using data obtained through the use of Spotipy. Spotipy is a Python library for Spotify Web API. It provides Python user a relatively easy way to scrape Spotify data legally. The scope of data obtainable through the API is limited. They are mostly catalogue information and current popularity of tracks, albums and artists. Apparently, there is no access to historical data. To better serve the purpose of the study, data regarding music consumption from other sources are cited. There will also be personal opinions and interpretation based on my knowledge on jazz.
Most Spotify data were retrieved on 16th February, 2022. Authorization flow used while connecting to the API was client credentials. I also tried using authorization code, which allows access to user information. There seemed to be no effect on search results returned. Data were visualized using Tableau. For the link to the Tableau workbook, please click here. Feel free to scrutinize how I manipulate the data.
Spotipy methods sp.search(), sp.tracks(), sp.albums() and sp.artists() were used to obtain catalogue information, popularity and URI of track, album and artist. Queries used to search for jazz tracks are “genre:jazz”, “genre:jazz fusion” and “jazz:bossa nova”.
The behaviour of the search endpoint of the API is very quirky. Small difference in the query like using “genre:jazz fusion” and “genre:jazz_fusion” will lead to different search results, sometimes vastly different. I plan to write a separate article to document it, together with the data collection processes and my experience using the Spotipy package.
This is part 1 of the study because I intend to compare the results of analysis of data obtained with a different search approach, and data of other music genres. There will be part 2 and potentially part 3 and 4.
Due to the nature of jazz, it is very difficult to define what it is and what it is not. Therefore, I will be loose on the definition of jazz intentionally in the following analysis. I also use fuzzy categorization on artists and tracks, e.g., artists that are highly-associated with jazz and artists that are not.
Short intro on jazz and its significance
Jazz is the child of the perfect marriage of the rhythm of West African music and the harmony of European classical music. Jazz is typified by swing feel and improvisation. Swing feel does not necessarily means swing rhythm. It can be other groove. In some cases, the regular pulse we perceived as rhythm is abandoned altogether. Similarly, improvisation is not always present, albeit rarely completely absent.
It has been over a century since its inception. The development of jazz was once suffered from strict racial segregation laws, yet it was able to unite black and white, musicians and audience, on and off the stage alike. Thankfully, it is still very much thriving today. It finds home in every continent and is now played and enjoyed by people of all races. In its heyday, i.e., 1920s to 1930s, it was literally the pop music of the era.
With all that being said, it is no longer as popular as it was. However, it is assimilated into other genres of music, just like it assimilates influence from other genres. Progressive rock musicians use jazz harmony. Fusion proponents improvise on rock rhythms. While Bossa Nova is a combination of samba rhythm and jazz harmony. Nowadays, the influence of jazz is literally everywhere. This is just the nature of it.
Data quality
The first thing that caught my attention is actually the quality of data, which also somewhat echoed with what I said in the previous paragraph. While it is impossible to go through all the tracks in the dataset to see if they are relevant, visualizing the number of tracks by artists gave me a pretty good idea of how the data looked like.
The above artists were grouped into two sub-categories: “Elevator Music” and “Others” (jazz-associated artists). “Happy Birthday All Names & Genres” and “The Bossa Nova All Stars” were the “artists” with the most tracks (almost 600 and 500 respectively). Together with the other four “artists”, they constituted about 32% of all tracks in the dataset. Unlike “The Bossa Nova All Stars”, which is not real musician (explained in below paragraphs), the name “Bossa Nova Covers” is a bit tricky. They are a real couple of musicians doing decent covers of classics in Bossa Nova style. “Kadri Gopalnath” is the odd one out here. Apparently, all his music is south Indian classical music (carnatic music). I guess Spotify associated him with jazz because he plays saxophone, which is a pretty lame reason.
Spotify seems to associate jazz with artists who are not normally associated with jazz by any stretch of imagination. I do not know how they categorize artists. I believe they do not rely very much on the artists genres tag to recommend music or create playlists since their suggestions work great for me 90% of the time.
Jazz as elevator music
The term “elevator music” is the closest description for the tracks these “artists” produced. The term elevator music refers to a kind of music that is meant to serve as background music rather than played for enjoyment and appreciation. It is often found filling sonic space of places like shopping mall, restaurants and bars.
The characteristics listed below led me to conclude tracks produced by those “artists” are elevator music.
- There is not a single jazz standard found after skimming through dozens of albums, probably to avoid royalties and copyright infringement.
- Many names of the tracks and albums suggest moods, venues, events or functions. For example, there is an album called “Background Jazz” with track names like “Peaceful Swing”.
- The timbre of the instruments sound suggests the tracks could be produced using digital audio workstation (DAW) with sound libraries instead of recording of performance by human.
- Most tracks lack the complexity typically heard in jazz. Some albums like “zZz Exclusive Lounge Music zZz” are filled with tracks that sound basically the same, i.e., same chord progression, tempo, dynamic and instrumentation, but a “saxophone soloist” playing short musical phrases that sound like random (or bad algorithms) rather than intentional.
- There are at least 10 more albums in similar names with the same tracks, and many more “artists” that do similar things on Spotify.
- In other words, they were produced with the lowest budget in the shortest period of time possible.
It is obvious that “Happy Birthday All Names & Genres” is not a real artist. For “The Bossa Nova All Stars”, it seems to be a name shared by multiple people or groups of people. While a few seems authentic, most of the albums use very generic names and album covers like this:
Jazz (from the easy-listening end of the spectrum) has been associated with elevator music for decades, and I believe it will continue to be. However, this is not the form of jazz that deserves more attention. After all, elevator music is meant to be background music. It also lacks artistic value.
Jazz might be more popular than we think
A few things to note here:
- Unsurprisingly, popularities of elevator music was much lower than music that is not.
- It seems jazz-associated artists do quite well, with medians of artist and album popularities at around 50.
- In terms of spread of data, artist and album popularity are more dispersed than that of tracks. This is because artist and album popularity are aggregated from track popularity.
However, when I looked closer at individual tracks, say the top 20 most popular tracks, jazz was not doing as good as it seems.
A few things to note here:
- Among the most popular tracks, most of them are not really jazz tracks.
- Non-jazz tracks are more popular than jazz tracks, which is expected.
- Popular jazz tracks tend to be newer than non-jazz tracks. Tracks released in and after 2000, which I consider relatively new, are highlighted in blue.
- 3 out of 4 popular jazz tracks are indeed Christmas songs. Their high popularity might be due to the fact they got played a lot during the last Christmas, which happened less than 2 months ago. Recent plays have higher weight than past plays.
Even though I categorized the tracks very loosely, there are still very little jazz tracks. You may think the 3 tracks by Norah Jones can be categorized as jazz, but I would argue “influenced by jazz” would be a more accurate description. The fact that her album Come Away With Me was supported by an array of jazz musicians, and was released on Blue Note Records, might complicate things a bit. In fact, Come Away With Me won the Grammy Award for Best Pop Vocal Album, and her tracks “Don’t Know Why” and “Sunrise” won the Grammy Award for Best Female Pop Vocal Performance. If you still do not agree, you may listen to Diana Krall’s album Live In Paris, which won the Grammy Award for Best Vocal Jazz Album in the same year as Come Away With Me, to get a better grasp of the sound of jazz.
To put things into perspective, let’s compare album and artist popularity of Norah Jones, Diana Krall (let me throw in Esperanza Spalding as well).
At this point, it might seem that jazz is not gaining traction in recent years, but it is not getting worse either. According to the music reports published by MRC Data (formerly Nielsen SoundScan and Nielsen Music Products), jazz maintains at least 1% share of total music sales volume from 2015 to 2021, which is similar to classical music. Although I have no knowledge on how they collect these data and how they categorize music, I think it is possible that some of the genre-blending music produced by jazz artists like Snarky Puppy is counted towards genres other than jazz.
Not enough followers for living jazz artists
Things to take note here:
- Among the top 20 jazz artists with the most followers, 11 of them are deceased artists while only 9 are living.
- Among the artists shown here, living jazz artists have less followers than deceased jazz artists both on average and in total.
Before explaining why this is a problem, let’s talk about what does following an artist in Spotify entail. According to Spotify, following an artist indicates preference for that artists, which helps making better recommendations. In other words, one would be more likely to see music by that artists or similar music from similar artists. They would appear in the playlists generated by Spotify, e.g. Release Radar, Daily Mix, etc.
The problem with following deceased artists is that deceased artists no longer have new output. Copyright of their music becomes their estates when they die. As a result, revenue generated (royalties) through streaming will only go to those who hold the legal titles of the estates, which are usually families of the artists. On the other hand, this source of income can be utilized by living artists to produce more music, which is all we want. This is not to say do not listen to deceased artists though. Respecting the tradition is very important in jazz. Every good jazz artists have good knowledge of the jazz greats that came before them.
This phenomenon somehow echoes with the increasing trend of catalogue music consumption in recent years. According to the music reports by MRC Data, U.S. catalogue music consumption increased from 61% in 2017 to almost 70% in 2021. In other words, only 30% of the consumption was current music.
Conclusion
There is no question that jazz is here to stay regardless of form, be it instrumental, vocal, fusion or elevator music. However, I do hope that, one day, jazz will be appreciated by a wider audience like in its heyday.
Ever since the advent of jazz fusion in 1960, more and more musicians are blending elements of jazz into other genres like pop, rock, hip-hop and even classical. Nowadays, it is almost customary to throw in a couple of jazz chords and phrases to spice up a song. You do not need formal training to recognize those chords and phrases, you just need to listen more jazz to get a hang of it.
While one can enjoy any genre of music on emotional level without prior knowledge or training (I listened to free jazz before I know what jazz actually is), being able to spot musical elements and their stylistic origins will definitely make the listening experience much more enjoyable on intelligent level. By having a deeper understanding of the music, artists and styles you like, you build a deeper relationship with them and would appreciate them more.
Epilogue — Jazz for Everyone!
I made a Spotify playlist of jazz that encompasses old and new tunes, vocal and instrumental, fusion to bebop. I try to include as many living jazz artists as possible, as well as crossover efforts by famous pop/R&B artists like Billie Eilish, Amy Winehouse, Lady Gaga and Norah Jones (yes, she did sing jazz tunes). There are jazzy covers of pop and rock classics as well. I believe these tunes can help to accustom one’s ears to jazz. Hope you enjoy!
Reference
2015 Nielsen Music U.S. Report
https://www.techbooky.com/wp-content/uploads/2016/01/Nielsen-2015-year-end-music-report.pdf
Nielsen Music Year-End Report U.S. 2016
https://www.nielsen.com/wp-content/uploads/sites/3/2019/04/2016-year-end-music-report-us.pdf
Nielsen Music 2017 Year-End Music Report U.S.
https://www.nielsen.com/wp-content/uploads/sites/3/2019/04/2017-year-end-music-report-us.pdf
Nielsen Music Mid-Year Report U.S. 2018
https://www.nielsen.com/wp-content/uploads/sites/3/2019/04/us-midyear-music-report-2018.pdf
Nielsen Music Year-End Music Report U.S. 2019 https://static.billboard.com/files/pdfs/NIELSEN_2019_YEARENDreportUS.pdf
MRC DATA Year-End Report U.S. 2020
https://www.billboard.com/wp-content/uploads/2021/01/MRC_Billboard_YEAR_END_2020_US-Final201.8.21-1610124809.pdf
MRC DATA Year-End Report U.S. 2021
https://mrcdatareports.com/wp-content/uploads/2022/01/MRC_YEAREND_2021_US_FNL.pdf
Grammy Award
https://www.grammy.com/
How Long Does Copyright Protection Last?https://www.copyright.gov/help/faq/faq-duration.html#:~:text=As%20a%20general%20rule%2C%20for,plus%20an%20additional%2070%20years.
