You Hold the Power

Discovering Music
with YOUR Tastes in Mind

Getty Images/Ringer illustration

Getty Images/Ringer illustration


It started with Spotify's
Discovery Playlist

The satification levels were inconsistent
week to week...

Therefore, I decided to explore the process of uncovering music by "acting" as a computer algorithm that would use users' musical interests to handpick new songs for consumption


Two Items for Analysis

1. Spotify's U.S. Top Song Charts
2. My own personal 2019 Spotify data
(self-curated playlists)


(the data visualization above, collected from Statistica,
demonstrates that 50% of the U.S. population listens to music on the daily)

In the Digital Age, we have permitted computers and algorithms to dictate choice. Many Americans, no the world, listen daily to music, so why not have the best music that fits you?
André Solomon

THE DATA

TechCrunch

TechCrunch

United States Spotify Data

The U.S. Top Songs for 2019 can be pubically accessed on Spotify’s website. It provides the user to view daily or weekly top charts in various regions in the world. Including the option to download the data through a CSV file.

  • Each weekly chart, starting from the beginning of January 2019 to November 11th, 2019, I had to merge into a singluar excel sheet to display all weeks together

Primary Observations

  • To be considered in the top rankings, each individual song must have been streamed over 1 Million times
  • The highest streamed songs were from reputable artists (Post Malone, Lizzo, Ariana Grande, etc.)
  • Songs less streamed (closer to the 1 Million mark) were from less common artists such as Gryffin
  • However, we must be mindful that humans gravitate to newly relased songs by popular artists. Which makes sense when top streamed artists release successful albums, EPs, and/or singles. Their name alone can generate streams. Yet, artists have purposefully produced albums with shorten song durations to recieve streams and dominate the charts

Spotify

Spotify

Top Streamed Artist - Post Malone

Top Streamed Artist - Post Malone

A Least Streamed Artist - Gyrffin (OC Weekly)

A Least Streamed Artist - Gyrffin (OC Weekly)

Data Visualizations

Spotify (Privacy Tab)

Spotify (Privacy Tab)

Soundiiz

Soundiiz

Sounddiiz

Sounddiiz

André Spotify Data

Spotify grants users to access/download their personal data including playlist information, streaming history, search queries, followers, payment information, etc.

However, I use a program called Soundiiz which would take my 2019 created playlists and form a CSV file that would display: Song Title, Album, Artists, Genre, Year, Date Added and Playlists Name

Primary Observations

  • Determining my favorite artists proved difficult because I had already made a personal endeavour to listen to new/non-mainstream artists. Therefore, my playlists contained a huge collection of different artists not particularly showing a pattern of favorites
  • The majority of my music on my playlists were released 2015 and beyond, with a few older outliers
  • On average each playlist contained 55 songs
  • The data visually presents the fact I listen to numerous genres (i.e. Dance Pop), however, many of the genres belong to a main group (Pop, Inide, R&B, etc.) and can be consolidated

Data Visualizations

Top Charts

Using the data visualizations above, it helped generate a concise picture of what musical interests are popular between the United States and myself. In addition, I had these stipulations upon gathering the data:

  • Top 5 Genres (focused on my data since it was provided in CSV files)
  • Top 10 Artists
  • Top 20 Songs (focused on U.S. Data due to provided streaming information)

United States

Top 10 Artists

  1. Post Malone (1,934,770,363 Streams)
  2. Billie Eilish (1,389,115,702 Streams)
  3. Ariana Grande (1,106,091,150 Streams)
  4. Juice WRLD (1,934,770,363 Streams)
  5. Khalid (1,934,770,363 Streams)
  6. Lil Nas X (1,934,770,363 Streams)
  7. Drake (1,934,770,363 Streams)
  8. DaBaby (1,934,770,363 Streams)
  9. Lil Baby (1,934,770,363 Streams)
  10. Lizzo (1,934,770,363 Streams)

Top 20 Songs

  1. Sunflower by Post Malone (362,077,524 Streams)
  2. Wow. by Post Malone) (282,212,779 Streams)
  3. 7 Rings by Ariana Grande (264,655,970 Streams)
  4. Middle Child by J.Cole (252,995,170 Streams)
  5. Bad Guy by Billie Eilish  (230,606,242 Streams)
  6. Going Bad by Meek Mill (226,424,862 Streams)
  7. Truth Hurts by Lizzo (208,935,149 Streams)
  8. Drip Too Hard by Lil Baby (192,943,760 Streams)
  9. Ransom by Lil Trecca (191,449,347 Streams)
  10. Swervin by A Boogie Wit Da Hoodie
    (185,454,579 Streams)
  11. Sicko Mode by Travis Scott (182,428,567 Streams)
  12. Murder on My Mind by YNW Melly
    (179,084,166 Streams)
  13. Thank u, Next by Ariana Grande (171,465,406 Streams)
  14. Without Me by Halsey (167,713,165 Streams)
  15. A lot by 21 Savage (165,750,174 Streams)
  16. Suge by DaBaby (161,562,082 Streams)
  17. Señorita by Shawn Mendes (158,646,946 Streams)
  18. When’s the Party Over by Billie Eilish
    (157,732,970 Streams)
  19. Look Back at It by A Boogie Wit Da Hoodie
    (157,516,676 Streams)
  20. Money in the Grave (Drake) (156,528,402 Streams)

André

Top 5 Genres

  1. R&B
  2. Pop
  3. Hip-Hop
  4. Indie
  5. Classical

Top 10 Artists

  1. Snoh Alegra
  2. Solange
  3. BENEE
  4. UMI
  5. DDG
  6. Inner Wave
  7. Pip Millett
  8. Lucky Daye
  9. Tame Impala
  10. Rex Orange County

Since the objective is to discover new songs for consumption, there will be limitations on the selected artists and songs using the critera above as reference:

  • Songs that have an older release date prior to 2018 will not be considered
  • Artists selected, nor their songs, cannot be on the top charts or within my own playlists
  • Chosen artists and/or songs must be derivatives (related) to Artists or songs listed above
  • Each playlist will only contain between 20-30 songs

*Light Green Highlight = A Picture to the Right*

GK Artogrpahic
Medium
Youtube

Taking the Data Further

The Process of Creating Playlists

Spotify

Spotify

Now that I have all this data,
what should I do with it?

I decided to use the top charts (Artists, Genres, Songs) generated by both the United States and I's Spotify data, alongside websites provided by Spotify web developers, to discover relatable artists that correlated with the data.


Artist Explorer

This tool helps you quickly discover related artists in beautiful graphical trees.


Discover Quickly

Quickly scan through
songs and save them
for later listening.


Magic Playlist

An intelligent algorithm developed under Spotify's API that enables users to create a playlist based on a song.

Creating the VIP List:
Cutting Down the Artists

As I gathered the final list of artists from the online resources, I had way too many artists (For the United States about 70 and myself 93). Therefore, these requirements guided my selection process:

For the United States

  • Artists had to have under 500,000 followers

For André

  • Artists had to have under 50,000 Followers

List of New Artists

United States

  • Cappadonna (65,619 Followers)
  • Carnage (251,311 Followers)
  • Cooliecut (71,505 Followers)
  • CupcakKe (342,601 Followers)
  • DaniLeigh (236,460 Followers)
  • Finneas (269,363 Followers)
  • Gianni and Kyle (107,234 Followers)
  • Girl in Red (494,868 Followers)
  • HoodRich Pablo Juan (254,072 Followers)
  • Internet Money (9,127 Followers)
  • Key Glock (381,695 Followers)
  • LeiKeli47 (114,791 Followers)
  • Lil Gotit (99,168 Followers)
  • Lil Keed (218,418 Followers)
  • Lost Kings (202,825 Followers)
  • Marc E. Bassey (270,547 Followers)
  • Mike Stud (327,741 Followers)
  • Mxmtoon (470,851 Followers)
  • Peewee Longway (278,318 Followers)
  • Qveen Herby (106,716 Followers)
  • Rico Nasty (300,805 Followers)
  • Stunna 4 Vegas (76,903 Followers)
  • Suecco the Child (36,786 Followers)
  • Tate McRae (295,993 Followers)
  • Todrick Hall (181,504 Followers)
  • Wifisfuneral (390,943 Followers)
  • Yung Tory (58,795 Followers)

André

  • Alex Isley (17,703 Followers)
  • Alextbh (38,578 Followers)
  • Amber Oliver (15,189 Followers)
  • Arlo Parks (20,779 Followers)
  • Ben Lukas Boysen (15,566 Followers)
  • BOYO (19,795 Followers)
  • Eliza (34,173 Followers)
  • Enjoy (28,248 Followers)
  • Free Nationals (38,995 Followers)
  • Hala (28,598 Followers)
  • Hot Flash Heat Wave (38,410 Followers)
  • IYAMAH (14,581 Followers)
  • Jean Deaux (31,703 Followers)
  • JGrrey (17,244 Followers)
  • Justin Jesso (13,135 Followers)
  • Kaiit (38,850 Followers)
  • Lava La Rue (15,664 Followers)
  • Marco McKinnis (13,348 Followers)
  • Mereba (43,868 Followers)
  • Nikolai Kapustin (4,293 Followers)
  • Rimon (27,299 Followers)
  • Shay Lia (15,929 Followers)
  • Symposium (23,555 Followers)
  • Tone Stith (41,259 Followers) 
  • Vacations (36,603 Followers)
  • Vansire (49,373 Followers)
  • Worn-Tin (9,514 Followers)

*Light Green Highlight = A Picture to the Left*

Zerchoo
Foundations Music
Atwood Magazine
Pianodao

Picking the Songs

I will conduct my song collections upon Spotify's Desktop Platform, which allows for easy access to create playlists.

Out of each new artist's 2019/2018 discography (including albums, EPs, singles, etc.) I will make my decisions based on Spotify’s popularity bar presented to the far right of each song. I will primarily favor picking from an artist’s albums so that potential listeners can later indulge in the entire album, promoting further investigation of the artist.

I will listen to the United States song selections to curate their playlist but for my own, I will only select the music, however, will not listen until my playlist is completed. Making it a true discovery playlist.

Final Deliverables

United States' Playlist

André's Playlist

Bringing it All Together

Top view of our Shorthand logo cake with dark ganache, gold sprikles and fresh berries.

Andriod Police

Andriod Police

Research

As I gathered data about the United States and myself alongside my skepticism upon Spotify's reccomendation engines, I wanted to look further into the backgrounds of what makes an artist or song popular and the reasons why a song might get plugged into our suggestions.

Questions I Had

  • Does Spotify’s discover feature incoprorate clout?
  • Are well-known artists given the opprotutnity to be “discovered” over others?
  • Are there monetary transactions that take place?

As I discovered in my research, many United States individuals were drawn to hits compared to other sources of music, but I have to agree with Knibbe that “...it’s not only due to a lack of time to explore the long tail, nor the fear of missing out, that is driving people to listen to hits; it’s also because social proof and familiarity are quite reassuring when it comes to uncover[ing] something new” (Knibbe, 2017). People like to discover new, but at the end of the day familiarity will prevail. Even myself, when a new album is released by Beyoncé, you better believe I will be listening that same day compared to an artist's newly released work that I am unfamilar with.

Streaming services take on the role of both taste-maker and gatekeeper. Platforms, such as Spotify, generate music selections that will cater to individual's interests, yet, coincidently put out music that fits a certain standard. For example, until recently (2018), Spotify was only working with record label artists in comparison to independent artists. Taking this in mind, artists that have somewhat established themselves have a greater opprotunity to be marketed by Spotify. Luckily, Spotify has started to slowly incorprate independent artists into the mix. However, independent artists still have a hard time because in order to be discovered they need to be placed in "popular" playlists either created by Spotify, popular users or well-known artists. Besides the artists Spotify works together with, the company also takes interest in their own created playlists (i.e Mood/Genre Playlist) and heavy followed artists. Furthermore, the computer algorithms that manipulate Spotify recommendations are then filled with bias, where new music is pushed upon users even with a plethora of selection. Even a computer sciencists code contains bias due their particular format choices.

Economic infuences are also intertwined with an artist's or song's popularity. It is not hard to comprehend that more familiar artists tend to be the most played. As they are popular artists overall, not just on Spotify, the two parties agree to promotional dealings where the partnership secures Spotify more users and the artists more attention. For example, you probably have seen an artist's face on one of Spotify's curated playlists. Not only is a famous artist supporting Spotify, but providing users the insight that the music contained in this particular playlist is noteworthy, therefore, you better listen. Basically, “their work is advertised prominently throughout the Spotify player in the ‘discover’ tab, on their main landing pages, or featured in ‘exclusive’ Spotify Sessions. These efforts represent an advanced promotional feedback loop that mixes user activity with interface design” (Morris & Powers, 2015, p. 113). But, simply artists have the capablity of working with spotify to create promotions (involves money) for their music, such as news upon an upcoming album release or concert.

THE FINAL CADENCE HAS COME



Takeaways

  • People are influenced by new releases
  • People of Color are controlling the charts due to the amount of music being released and more accessible channels
  • People default to "hits" because humans are drawn to what is familiar
  • Spotify gives weight to their own playlists and artists with more followers
  • Bias controls what people listen to, including algorithms and this project
  • Artist Promotions ($) = More Followings

Call-to-Action

I implore music listeners to map their musical interests to determine what music alligns to them personally and take advantage of the resources Spotify provides.

  • Do not just let computer algorithms curate you music selections but if you going to let Spotify do the work, there are ways to help influence improved results, such as: Click the “heart” icon next to the song, follow your favorite artists, add songs to your own personally curated playlists, skip songs you do not like quickly and try not be a passive listener.

Wanna Do This? (Grab Your Data)

Gathering Common Spotify Data

  1. Navigate to https://spotifycharts.com/regional to download data from Spotify’s Top Charts
  2. Select Top 200 or Viral 50 depending on the purpose of your analysis—to see top tracks or trending tracks
  3. Use the drop-downs to filter by your specific country or select Global to see all country data
  4. You can also choose to select Daily or Weekly data
  5. Click “Download to CSV” in the top right-hand corner (Note: That the downloaded CSV file will only generate a single day or a single week, therefore, bigger datasets will require you to combine the files. My method was through excel)

Gathering Personal Spotify Data (There are two ways of doing this)

  • Using Last.fm (a music service that lets you track your music with what they call “scrobbling.”) and connecting to Spotify account (or other music streaming services) —> Wish I knew about this method earlier 
  1. Create a free last.fm account
  2. Enable scrobbling so that Last.fm can track all your listening across your digital music services (including Spotify)
  3. Once Last.fm has a few weeks of listening data, you can use this Last.fm to CSV converter —simply use your username and it’ll create a CSV for you.
  • Download Personal Spotify Data (only provides a three-month snapshot and can be tricky to decipher)
  1. Create a free/premium Spotify account or Login to an established account
  2. Once logged into Spotify, head to the privacy page where there is an option to download your data.
  3. Wait a few days for Spotify to send you an email with your data.

Although this is an easier method, you only get a three-month snapshot of data. But there’s still plenty in there to get some great insights into your recent listening habits.

See Behind the Scenes

Visit my Github page which explains a deatiled process of this project

Spotify

Spotify

References

  • Knibbe, J. (2017, March 07). Is Music Personal? – Deezer I/O. https://deezer.io/is-music-personal-315106ec0179
  • “MagicPlaylist.” MagicPlaylist. Accessed November 2019. https://magicplaylist.co/#/?_k=n8gf2x.
  • Morris, J. W., & Powers, D. (2015). Control, curation and musical experience in streaming music services. Creative Industries Journal,8(2), 106-122. doi:10.1080/17510694.2015.1090222
  • “Discover Quickly.” Discover Quickly. Accessed November 2019. https://discoverquickly.com/.
  • “Spotify Artist Explorer.” Spotify for Developers. Accessed November 2019. https://developer.spotify.com/community/showcase/artist-explorer-spotify/.
  • “Spotify Charts.” Charts. Accessed November 2019. https://spotifycharts.com/regional.
  • Soundiiz. “Transfer Playlists and Favorites between Streaming Services.” Soundiiz. Accessed November 2019. https://soundiiz.com/.
  • “Music Listening Habits in the U.S. by Age 2019.” Statista. Accessed November 2019. https://www.statista.com/statistics/749666/music-listening-habits-age-usa/.

Images

  • Getty Images, and Ringer Illustrations. “Spotify Volume Knob.” Can Spotify Solve the Art-vs.-Artist Problem?, The Ringer, Nov. 2019, https://www.theringer.com/music/2018/5/12/17346624/spotify-r-kelly-xxx-tentacion.
  • Tech Crunch. “Spotify Watching Users.” Spotify Needs to Crack down on Labels Snatching User Data, Tech Crunch, Nov. 2019, https://techcrunch.com/2019/06/27/cambridge-anamusica/.
  • https://commons.wikimedia.org/wiki/File:Post_Malone_Stavernfestivalen_2018_(202948).jpg
  • Billboard. “Gryffin.” Listen to Gryffin's 10 Best Remixes, Billboard, Nov. 2019, https://www.billboard.com/articles/news/dance/8495573/gryffin-best-remix-list.
  • Reddit. “Billie Eilish.” Billie Eilish, Aquarelle Coloured Pencils, A4, Reddit, Nov. 2019, https://www.reddit.com/r/Art/comments/a076k6/billie_eilish_aquarelle_coloured_pencils_a4/.
  • Medium. “Lizzo (Sailor Moon).” How Lizzo Keeps Me Sober, Medium, Nov. 2019, https://medium.com/@ericafreedman/how-lizzo-keeps-me-sober-4ba1419ce9a0.
  • IRIS DENG. “R&B.” How Rap, R&B, and Lo-Fi Artists Are Crossing the East-West Divide, The Varsity, Nov. 2019, https://thevarsity.ca/2018/01/22/how-rap-rb-and-lo-fi-artists-are-crossing-the-east-west-divide/.
  • “Snoh Aalegra in Head Towel.” Snoh Aalegra-In Your River, Youtube, Nov. 2019, https://www.youtube.com/watch?v=ut8zKDIT0f8.
  • “Music for Everyone.” Spotify, Nov. 2019, https://www.spotify.com/us/
  • Billboard. “Internet Money.” How Producers Internet Money Took Over the Hot 100 By Making Beats That Sound Like Other Artists, Billboard, Nov. 2019, https://www.billboard.com/articles/columns/hip-hop/8468428/internet-money-producers-interview.
  • Foundations Music. “Mxmtoon.” MXMTOON, Foundations Music, Nov. 2019, https://foundationsmusic.com/mxmtoon/.
  • Atwood Magazine. “Vansire Outside.” TODAY’S SONG: VANSIRE MUSE ABOUT MODERN LIFE WITH “METAMODERNITY,” Atwood Magazine, Nov. 2019, http://atwoodmagazine.com/vsmm-vansire-metamodernity-song-review/.
  • Pianodao. “Nikolai Kapustin Resting on His Hands.” Exploring the Piano Music of Nikolai Kapustin, Pianodao, Nov. 2019, https://pianodao.com/2019/07/26/exploring-the-piano-music-of-nikolai-kapustin/.
  • Spotify Playlist Music , radio PNG clipart
  • “MagicPlaylist.” MagicPlaylist. Accessed November 2019. https://magicplaylist.co/#/?_k=n8gf2x.
  • “Discover Quickly.” Discover Quickly. Accessed November 2019. https://discoverquickly.com/.
  • “Spotify Artist Explorer.” Spotify for Developers. Accessed November 2019. https://developer.spotify.com/community/showcase/artist-explorer-spotify/.
  • “Spotify Charts.” Charts. Accessed November 2019. https://spotifycharts.com/regional.
  • Soundiiz. “Transfer Playlists and Favorites between Streaming Services.” Soundiiz. Accessed November 2019. https://soundiiz.com/.