Like a present on your birthday, your 2024 music algorithm is wrapped tight with a bow. For many Spotify users around the world, the end of November and the beginning of December gives us one more reason to celebrate. Now that we can carry millions of songs in our pockets daily, we have all the more reason to see an accurate analysis of our yearly music summary. But how exactly does Spotify gather this data and adjust our algorithms to our ever-changing taste in musical vibrations?
As a global company and streaming platform, Spotify reaches more than 180 countries. Publicly released in 2008, the service has overtaken every other method of listening to music. With the goal of a “frictionless music experience for listeners,” Spotify employees dedicated countless hours toward receiving quality customer feedback in relation to desirable streaming preferences. This is what has made Spotify the domineering figure out of all other streaming platforms, taking over the music world.
Spotify officially titled Wrapped as their end-of-the-year music analysis for users in 2016 but didn’t become popularized until the feature was added to the app. Before 2019, users could only access their personalized year in music through the website. However, what truly revolutionized Wrapped was intern Jewel Ham’s story concept that inspired interactivity between listeners and different features. To dive deeper into what Spotify Wrapped truly is, we’ll observe the summary analysis, the 2024 popular music predictions, and how algorithms adapt throughout the year.
Employees are constantly updating users’ Wrapped summaries with different widgets each year. In 2023, Spotify listeners received a personalized character card associated with their music (“Me in 2023”), their perfect city match (“SoundTown”), along with their top-five most listened-to genres and artists, and a top-five artist’s message. They also get an in-depth data analysis of the number of artists and genres listened to, and how many minutes were spent in total.
Unsurprisingly, correlated with one of the most successful tours of our century, Taylor Swift was the most listened-to artist last year. She was followed by Bad Bunny, The Weeknd, Drake and Peso Pluma sharing the podium. In Rowan Davies’ article for TechRadar, he expects to see similar trends from last year with Taylor Swift landing in the top five alongside Troye Sivan, Charlie XCX, Sabrina Carpenter and Chappell Roan. Many of these artists’ recognition soared for varying reasons like virality on TikTok, hit album releases and popular collaborations. However each of these artists reached greater fame, it’s a strong prediction that they’ll top the Spotify charts in Wrapped.
Although there were rumors and concerns over the unofficial ‘deadline’ for Spotify to stop analyzing trends after October 31, Spotify X denounced these presumptions with a treat. However, they continue the mystery of the ambiguous deadline for Wrapped data to begin summarizing. The silver lining is that there’s no reason to stress over completely adjusting your Wrapped results on Halloween.
But that means throughout the year, Spotify will be tracking your regular music listening habits and suggesting new music based on your personalized algorithm. The Bandits for Recommendations as Treatments is an AI system designed to recommend previously listened-to songs as well as new ones with similar styles. It solves each user’s algorithm trend with a similarity/distance formula like the one displayed below:
This specific equation is known as the Euclidean formula used to calculate the direct distance between two points—in this case, the distance between yours and others’ preferences in music. Each person’s preferences are analyzed through the different playlists they create, the songs they skip (and how long one spends listening to a song before skipping), and where the location they listen from.
Consider the songs you listen to. Spotify will then categorize your music taste into four broad groups, clustering artists and music into genres: country, jazz, classic rock and folk. Each of your songs will be placed in these categories, spaced according to how extremely a song represents a specific genre. This is considered a t-SNE algorithm to map every song into a two-dimensional latent space.
After analyzing the audio data of the songs people listen to, The BaRT uses the 30-second rule to infer a listener’s enjoyment of recommended music. If they listen for longer than thirty seconds before skipping, Spotify will continue suggesting similar songs. The same goes for how much time one spends on a playlist in particular: the longer one listens, the more Spotify will recommend songs of a similar category.
Finally, Spotify takes into account the personal information that users share such as location, language and age. These demographics play a clear role in understanding what’s preferred or not preferred by a specific local, speaker or generation. For example, if French is your chosen language on Spotify, the algorithm will provide you with more French music recommendations.
Even though we have an influential role in the music we listen to—and what will contribute to our Wrapped each year—Spotify is constantly tracking our preferences with data analysis tools. By using collaborative, content-based, and language filtering, Spotify introduces us to old favorites and fresh songs we might not have listened to before. As the cherry on top, at the end of the year, our customized Wrapped profiles encourage us to set new personal records with our listening habits.
As a Spotify user, you must be eager to see your 2024 music trend analysis. Now that you’ve learned a great deal about the complex systems and programs Spotify uses to track your preferences, you just might enjoy this present a little more. Spotify is your personal music assistant, supplying your music taste with all its necessities and more. So while you’re listening to a new song on a bus ride or on your way to class, remember that Spotify sent you this music to enjoy—but don’t be afraid to skip before thirty seconds is up!
