In this project, I set out to understand something that has irked me for years (why are the song credits on Spotify (introduced in 2018) so lousy?) and communicate this problem to others through a practical example.
What I found out was that this is mainly a data collection and storage isssue. Artists are responsible for getting accurate credit information to their record label, who is responsible for sending to streaming services, who are responsible for entering it in and displaying it. However, this process hasn't been systematized, so there will be many cases where the information is missing or incomplete. Moreover, the information that is displayed is very rudimentary and not particularly helpful for anyone who wants to see all the collaborations a given artist has done.
For the case study, I chose Frank Ocean's 2016 album, Blonde, both because of the number and breadth of artists he included in the project, and because the tiny number of credits that Spotify displays (see spotify-screen-captures). I determined the true credit information from pictures of the physical CD which agree with the information available on Wikipedia. As an extra step, I also included the credits for the samples used in Blonde, which interestingly enough is not on Wikipedia yet.
To analyze the data, I simply created a list of all individuals credited on the album, and pointed out people who I found to be notable because they were famous artists, up and coming artists, or older artists. I chose these categories because I thought they corresponded to what Frank Ocean fans might see and go "Hey! I should give that person a listen."
This project was my first forray into scrollytelling using Scrollama and D3. I am pretty happy with the first transition, which is sort of this grand reveal of how much info the Spotify credits are missing. If I were to revist this, I would like to dive a little deeper into the analysis and break the credits up by track, type of credit (writing, produciton, musican, vocals, arrangement, technician). For this, I would probably want to start with a more sophisticated data store (namely D3) as I just whipped up the existing graphic using Adobe Illustrator and ai2html. Adding the ability for desktop users to mouseover and see full credit information would also be a nice-to-have.