Mapping New York’s Local Music Scenes
Using Bandcamp data to see which genres are the most popular in NY’s localities
Back in the day of my angst-y teenage years, playing in pop-punk bands and going to local shows, you’d often find band members, or their fans, handing out a homemade demos on CD-Rs. Those demos were most likely recorded on a 4-track tape recorder (or digital if you were fancy). Then Myspace came out my senior year and completely changed the way the local scene shared and promoted music. Rather than handing out a CD, you could give someone a card with a link on it to go listen to your music. Today it is way cheaper and easier for a musical artist to record their songs and publish them online. You can pay digital distribution services such as DistroKid and TuneCore, or you can put your music on SoundCloud or Bandcamp for free.
Out of all these services, Bandcamp is the best place for an independent artist to make money from of their recordings and merch; they give artists between 80–85% of your purchases. Being a fan and artist on Bandcamp, and a lover of data, I created a dataset on Qri of Bandcamp 79,000 artists that includes the artists’ names, genres, locations, geo coordinates, and Spotify data (if they are on Spotify). With that dataset, I made a map of what New York state’s local scenes look like by finding what genre is the most common.
The approach
I found that Bandcamp’s official API is limited to the use of artists and labels in order to manage their accounts, however, they do have an artist index page where I was able to collect 1,861,236 artist names and URLs. With that information, I scraped what I could of the URLs to collect the genres and location of each artist. I then used GeoPy to geocode the locations so I could place them on a map. I also used Spotify’s API to find if the artist’s music was available there, and collect the number of Spotify followers they have and their Spotify popularity score. That part was a bit tricky, because there are so many artists who share the same name. I explain in more detail how I tried to keep the data consistent in the dataset’s readme on Qri Cloud.
Local Scenes of New York
We created a web map of each city’s artists, with links back to their bandcamp pages!
🗺 View the Interactive Web Map Now 🗺
There were 179 cities/towns in New York state with a total of 3,149 artists. With New York City, the boroughs all have a different scene happening. In Manhattan, Electronic music is the most dominant, while the Bronx is still home to Hip-hop. Brooklyn and Staten Island have more Alternative artists, and Queens has more Hip-hop. In other parts of the state like Ithaca, Experimental music is their jam, and in the state capital, Rock music is still alive. The code for the map of New York can be found on Github.
Interesting notes
As I was trying to collect the genres and locations of those artists, I found that there are a lot of fake or empty pages. Some of those fakes are for The Weeknd and Demi Lovato. However, there are popular independent artists on Bandcamp as well like Chance The Rapper. Another thing I noticed was the distribution of genres. Genres are tags on Bandcamp and they are defined by the user, so there are a lot of unique ones (over 21,000). Some of those genres include “plunderphonics”, “breakfast”, and — my personal favorite — “lit”. To cut down on the number of unique genres, I used Bandcamp’s tags page as a reference to the most popular genres. The majority of artists in America that are on Bandcamp are tagged as either “alternative” or “electronic”.
Versioning and Publishing the dataset with Qri
There are plenty of more artists to add to the dataset, along with finding and verifying their music on Spotify to add another feature to the data. More complex maps could be made as well, specific to different regions. With Qri, you can clone a copy of the dataset for yourself, and put together a map of your state or country’s local scene. But maybe the most updated version of the dataset isn’t what you need. That’s ok, because Qri allows you to clone previous versions of the dataset, along with viewing the differences between the versions as well. You can find the dataset here to get started.
Thanks for reading!
