I decided to conduct an experiment in 2018 and embrace the coming age of the bots. Throughout the year, I culled songs from Spotify’s Discover Weekly1 recommendations, music heard in the wild (thanks Shazam), songs surfaced by Apple Music, and tracks played on local radio (esp. KUTX).
I say that I embraced the coming age of the bots because well over 80% of the ~1,500 - 2,000 songs I listened to (and tracked) during the year were purposefully based on algorithm-generated recommendations. Of these, I saved 228.2 I then filtered for only those songs released in 2018, reducing the count to a reasonably compact 45 (clocking in just under 3 hours of total play time). And in one final nod to letting the computers do the thinking, I sequenced the song order on “shuffle”. I’ve published the final playlist on Apple Music.
It should be noted that I was inspired to share these results by a friend of mine who regularly posts his favorite songs and albums of the year. Interestingly, though a couple of artists (CHVRCHES, Courtney Barnett) found their way on both of our 2018 lists, not a single song was duplicated.
As a point of comparison, I also installed Federico Viticci’s Siri Shortcut Apple Music Wrapped, which attempts to capture for Apple Music customers the spirit of Spotify’s year-end listening trends summary. The 25 songs that comprise my resulting "Wrapped (2018)" playlist are not limited by year of release, as the experiment above, but are selected solely based on play count. As Apple further embraces services, one can hope they will bake-in these kinds of features in the future.
1. Spotify serves up 30 songs a week to “discover” based on an algorithm which assesses your listening habits, saved songs, and, from what I have read, songs others on the network are sharing, saving, and what not.
2. 90%+ of the down-selected songs originated with Discover Weekly.