Spotify on Webpages

Project Overview

Inspired by (Muir 2024) who is a MEDS alumni, I want to extract and display data from my Spotify. Similar to the annual Spotify Wrapped information like my top tracks played and top artists can be retrieved. Additionally, I wanna try to get track audio features (track danceability, energy, loudness, key, etc.) like Sam did in their project.

Using the {spotifyr} package

The {spotifyr} package requests data from the Spotify API and we are able to load this data into R.

# Load packages
library(spotifyr) #API interaction
library(tidyverse)
library(kableExtra)
library(httr) # Troubleshooting redirect url
source(here::here("posts/2025-11-18-spotify/keys.R")) # My API keys

If you’d like to follow along, do this set up:

1. Create a Spotify Developer Account on the Spotify Developer Website.

2. Create an app with a title, description, and redirect url.

• As of April 9th, 2025, localhost string is not allowed in Spotify redirect URI’s. See this {spotifyr} GitHub Issue. I set the redirect as http://127.0.0.1:1410/ and I followed the solution in the GitHub Issue.

3. Save your client ID & client secret.

4. I sourced these keys from a separate R script to keep them private. The following chunk is how to set up the keys as environmental variables in R.

# Setup client ID & Secret in keys.R
Sys.setenv(SPOTIFY_CLIENT_ID = SPOTIFY_CLIENT_ID)
Sys.setenv(SPOTIFY_CLIENT_SECRET = SPOTIFY_CLIENT_SECRET)

# From GitHub Issue, re-define authorization function
get_spotify_authorization_code_fixed <- function (client_id = Sys.getenv("SPOTIFY_CLIENT_ID"), client_secret = Sys.getenv("SPOTIFY_CLIENT_SECRET"), 
    scope = scopes()[c(1:19)]) # Trying Sam's scopes in newly defined function
{
    endpoint <- oauth_endpoint(authorize = "https://accounts.spotify.com/authorize", 
        access = "https://accounts.spotify.com/api/token")
    app <- oauth_app("spotifyr", client_id, client_secret, redirect_uri = "http://127.0.0.1:1410/")
    token <- (purrr::safely(.f = oauth2.0_token))(endpoint = endpoint, 
        app = app, scope = scope)
    if (!is.null(token$error)) {
        token$error
    }
    else {
        token$result
    }
}

Now running the fixed function should work to get an authentication code.

auth_code <-  get_spotify_authorization_code_fixed()

Use auth_code generated from the fixed authorization code function and we can now call (some) {spotifyr} functions!

# Define function to extract genres from genre list column
create_genre_string <- function(genre_list) {
  paste(genre_list, collapse = ", ")
}

# Get top artists
top_artists <- get_my_top_artists_or_tracks(type = "artists", 
                                            limit = 10,
                                            time_range = "medium_term",
                                            authorization = auth_code) |> 
  mutate(genre_string = map_chr(genres, create_genre_string)) |> 
  select(c(name, genre_string, popularity, followers.total))

# Define function to extract artist names from artist dataframe column
create_artist_string <- function(artist_df) {
  paste(artist_df$name, collapse = ", ")
}

# Get top tracks
top_tracks <- get_my_top_artists_or_tracks(type = "tracks", 
                                            limit = 10,
                                            time_range = "medium_term",
                                            authorization = auth_code) |> 
  mutate(artist_list = map_chr(artists, create_artist_string)) |> 
  select(c(name, artist_list, album.name, album.release_date))

Other cool functions!

# Get current playing track
current <- get_my_currently_playing(authorization = auth_code)

# Get recently played tracks
recents <- get_my_recently_played(limit = 3, authorization = auth_code)

Since I had to source the keys I can’t produce a rendered, up to date output. But I’ll come back to this post, re-run everything, and see how things change. So instead, I have to export the R objects as a csv and read them in.

# # Export top tracks as csv
# write_csv(top_tracks, file = 'posts/2025-11-18-spotify/zach_tracks.csv')
# # Export top artists as csv
# write_csv(top_artists, file = 'posts/2025-11-18-spotify/zach_artists.csv')

# Read in my csvs
zachs_tracks <- read_csv(here::here('posts', '2025-11-18-spotify', 'zach_tracks.csv'))
zachs_artists <- read_csv(here::here('posts', '2025-11-18-spotify', 'zach_artists.csv'))

zachs_tracks |> 
  kbl(col.names = c("Track Name", "Artists", "Album Name", "Album Release Date"))

Track Name	Artists	Album Name	Album Release Date
Hey you ?, Hey sup ?	Masayuki Suzuki	Snazzy	2024-03-27
Red Swan	YOSHIKI, HYDE	Red Swan	2018-10-03
Psychedelic City	Masayuki Suzuki	Snazzy	2024-03-27
YOU	JOSS, Kenny Pham, KELIZA	YOU	2025-04-11
Full of Grace (I Refuse to Tend My Own Grave)	$uicideboy$	THY KINGDOM COME	2025-08-01
Sakura Biyori and Time Machine with Hatsune Miku	Ado, Hatsune Miku	Ado's Best Adobum	2025-04-08
Good Things Fall Apart vs. Sad Songs (With Said The Sky feat. Annika Wells)	ILLENIUM, Jon Bellion, Said The Sky, Annika Wells	ASCEND (Tour Edits)	2020-06-12
Whiplash	Crankdat, SOFI	Whiplash	2025-03-14
風のゆくえ	Ado	ウタの歌 ONE PIECE FILM RED	2022-08-09
Ichiban no Takaramono - Yui final Ver.	LiSA, VISUAL ARTS / Key	Ichiban no Takaramono -Yui final ver.- / Girls Dead Monster STARRING LiSA	2010-12-08

My Top 5 Tracks

I’m displaying the above Spotify content manually through the Creating an Embed feature from Spotify.

zachs_artists |> 
  mutate(followers.total = scales::comma(followers.total)) |> 
  kbl(col.names = c("Artist Name", "Genres", "Popularity", "Total Followers"))

Artist Name	Genres	Popularity	Total Followers
$uicideboy$	emo rap, horrorcore, cloud rap, trap metal, underground hip hop	84	9,006,324
Masayuki Suzuki	anime, j-pop, kayokyoku, city pop	51	315,706
Ado	j-pop, anime, vocaloid, j-rock	76	6,790,414
ILLENIUM	melodic bass, future bass, edm, techno	71	1,635,229
Sawano Hiroyuki	anime, soundtrack	66	733,386
LiSA	anime, j-pop, j-rock	69	3,618,103
Ray Volpe	dubstep, riddim, deathstep, melodic bass, bass music, edm	53	173,553
Crankdat	dubstep, riddim, deathstep, edm	53	200,462
SLANDER	melodic bass, dubstep, future bass, edm	64	594,841
Seven Lions	melodic bass, future bass, edm, chillstep, dubstep, techno, progressive trance	59	523,928

My Top 5 Artists

Conclusion

Unfortunately, numerous endpoints in the Spotify API were deprecated in November 2024 outlined in this Spotify for Developers Post, including Get Track’s Audio Features (for danceability, energy, etc.). This was due to an update in Spotify’s policies where they no longer wanted users to input these data into a machine learning/AI model.

However, this was still a cool project to do a little bit of wrangling in R and also a slight introduction to APIs for me. Again, I’ll be re-visiting this blog post and updating it periodically to see how my Spotify stats change. Thanks for tuning in!

References

Muir, Sam. 2024. “Who’s on AUX? Using Machine Learning to Predict Whose Spotify Playlist a Song Belongs To.” April 13, 2024. https://shmuir.github.io/projects/2024-04-13-spotify-predictions/.