Adele is an Amazing Singer-Data can Show Why!
If you’re trying to prove the awesomeness of your favorite singer, why not use data?
Songs have been around for ages, and have inched their way into humans’ lives, one way or another. I started regularly listening to songs when I was 9 or so, and nearly a decade later I’ve immersed myself in many different genres of songs.
Out of all the artists I’ve listened to over the years, my favorite by far is Adele. If you’ve listened to her most famous songs like “Hello”, “Someone Like You” or “Rolling in The Deep”, you would know that her vocals are rich and mesmerizing, and her songs really pour their strong emotions into your heart. They make you feel things you’ve never felt before.
Listening to “Someone Like You”, made me feel heartbreak for the first time when I was 10.
As part of a data science project, I decided to make use of my appreciation for songs and Adele, and analyze this dataset. It is a collection of audio features for 160k+ tracks released in between 1921 and 2020, on Spotify.
Spotify is a digital music, podcast, and video streaming service that gives you access to millions of songs and other content from artists all over the world.
The result?
I came up with data-driven reasons for the question: Why is Adele is an amazing singer?
The Main Supporting Questions
In order to properly show Adele’s amazingness, I needed to ask some supporting questions and find their answers within the dataset.
Here are 3 questions I had:
- Are there any notable recent trends in songs?
- How popular is Adele compared to her competitors in 2020?
- What are the key differences between Adele and her top competitors?
In the following sections, I will only discuss parts of my findings that are relevant to my arguments. If you want a complete set of findings, you can check out my Kaggle notebook or my github.
More on the Data
Before we dive into the answers, I’ll explain some of the relevant but not-so-straightforward features of the dataset first. Feel free to scroll back here if you are unsure what a feature means.
valence: How musically positive/happy a track is.popularity: How popular a track is at the time of obtaining the data. It is important to know that the dataset was released in June 2020, so all the popularity features we use here, refer to the track’s popularity in June 2020.explicit: Whether or not a track contains explicit content. 1 for explicit, 0 for not explicit.speechiness: Levels of spoken words in a track. Closer to 1 means exclusively contains speech only. Closer to 0 means exclusively contains music only.
1. Song Trends in Recent Years
For this part, I framed “recent years” as from 2010 onwards. In other words, the observations I made, were derived from features of the average song in each year from 2010 to 2020.
Additionally, I also obtained another table of averaged features for the top 100 popular songs in each year from 2010 to 2020.
I got some pretty interesting findings in this part here.
a) Songs were getting progressively sadder from 2010, but have been trending to be happier starting 2017–2018
- The valences and energies for regular songs mostly dipped from 2010 all the way till 2017–2018, before spiking upwards again, meaning songs were getting musically sadder and lethargic, before bouncing up.
- Popular songs from 2010–2020 also follow this trend, reaching the saddest and least energetic levels in 2016–2018.
- The trends might suggest that 2016–2018 was a period full of heartbreak for the world, such that more songs were sadder, and these solemn songs ended up becoming popular. However, we need to do more data digging to prove this.
- Interestingly enough, 2020 seems to have encouraged songs to spike greatly in happiness and positivity.
With all the negativity in 2020, I guess the world turned to songs for a positive break.
b) Songs are getting more danceable.
Danceability shows an increasing trend for average songs and popular songs.
c) Songs are getting more explicit.
The average song is more likely to be explicit as the years pass, and such explicit songs are getting more popular.
d) What do the trends have to do with Adele?
At this point, you might be wondering, what is the relevance of these trends with Adele? To put it simply:
Adele is not “trendy”.
Take a look at the figures below:
Adele does not follow the trends of recent songs. Her songs have a low average valence compared to 2020 songs, as well as a low average danceability compared to 2020 songs.
In addition to this her explicitness is only 1/10th of that of 2020 songs:
With all of these features trending upwards from at least the last 3 years or so, it’s likely Adele will fall even further behind the trend if she sticks to her style.
Conclusion for Section 1:
Songs have been trending to become happier since 2017–2018, more danceable and more explicit. But Adele’s musical genre is within the realm of Pop, R&B and Soul. Hence, her songs tend to be more like emotional ballads full of deep meaning — they don’t follow happy, upbeat trends.
As a result, it’s easy to come to the assumption that Adele’s songs are probably not so popular in recent times. Well, that brings us to the next question.
2. Adele’s Popularity compared to her competitors in 2020
With the data we have, finding how popular Adele is in 2020 is straightforward. The track features already include popularity, so to get a good proxy for how popular Adele is, we can just sum up the popularities of all her songs, and compare this to the summed popularities of all other artists.
Doing this, we get :
This number looks reasonable, but we have to remember that the dataset contains many old tracks as well, from the 20th century. Meaning, when we use this method, we are comparing Adele’s songs to legendary classics from the likes of Frank Sinatra, Elvis Presley, etc.
In fact here are the top 5 most popular artists based on this method:
It seems unfair to call The Beatles or Frank Sinatra, “competition” for Adele. They are after all, timeless legends, and their songs have been known worldwide for decades now. Meanwhile Adele’s first song, Hometown Glory, was only released in 2007.
Therefore, I took a subset of artists who have released songs on 2007 onwards, and deemed those artists to be Adele’s competitors instead.
Here are the results:
From this point onwards, when we discuss the top 5 artists, we refer to artists that have released tracks on 2007 onwards.
Conclusion for Section 2
Adele is extremely popular in 2020. She is in the top 1.63% of Spotify artist worldwide, and in the top 0.3% of artists who have released tracks on 2007 onwards.
With this in mind, I sought out to find the answers to the last question.
3. Key Differences between Adele and her Competitors
Knowing Adele’s popularity, I wanted to find further what puts the top 5 most popular artists apart from Adele.
Is she lacking in some areas? Are there any common things between the top 5 artists?
With more manipulations, I found each of the top 5 artists’ average track features. In comparison to Adele’s average track features, these are the results obtained:
a) Their songs are happier, more energetic, and people are more likely to dance to their songs
- Of the 3 features energy, valence and danceability, all the top 5 artists’ average tracks are higher than Adele’s average tracks.
- In other words, the top 5 artists follow the recent trend in popular songs more than Adele, in terms of having happy and uplifting song melodies.
b) Most of the top 5 have songs with a higher spoken word ratio
- Drake, Future, Lil Uzi Vert, Lil Wayne all have an average speechiness that is at least 2.5 times as much as Adele’s average speechiness.
- This is expected, because with the exception of Taylor Swift, the rest of the artists in the top 5 are rappers.
c) Most of the top 5 artists have released many more explicit songs
- Once again, all of the top 5 artists except Taylor Swift, have a much higher explicitness than Adele (more than 10x difference).
- One might conclude that because they are modern rappers, they have a lot of explicit tracks out. Well, one important thing to note is that correlation is not the same as causation.
- In other words, just because they are all modern rappers, and they have high explicit content, it doesn’t mean that because of the former, the latter follows.
d) Interestingly enough, Adele’s songs have a higher average popularity than most of them.
- Adele’s songs are actually slightly more popular on average, compared to Future, Lil Uzi Vert, Lil Wayne and Taylor Swift.
- Her songs are also close in popularity to Drake’s songs.
So why is Adele not in the top 5 artists?
e) What Adele lacks
Well, since we summed up each of the artists’ songs’ popularities as a proxy for how popular an artist is, this leads to one possibility — Adele has released very little songs compared to the top 5 artists.
Sure enough, after more data exploration, we can see that :
Note: Here I use the words “has been involved”, because as long as a Spotify song has listed an artist as one of its creators, I included the song.
Conclusion for section 3
Adele’s music is relatively different in style and mood compared to those of the top 5 artists. However, this is not the reason why Adele is not in the top 5.
Instead, it is simply because she’s been involved in making very little songs compared to the top 5 artists.
Overall Conclusion : So why is Adele a great singer/musician?
The reason is a compilation of the 3 conclusions we had before.
1. Adele’s musically sadder, slower, more soulful genre of music does not follow recent popular song trends of happy and upbeat songs.
2. However, data shows that she is still within the top 0.3% of most popular artists (with tracks released on 2007 onwards).
3. In fact, the reason why she isn’t in the top 5, isn’t because of her non-trendy style, but rather because she has released little songs. The quality of her tracks speak for themselves, as their average popularity rivals the tracks released by the top 5 artists.
Now that we’re at the end, I’d like to pose questions for you:
When presented with these arguments, would you agree that Adele is a great musician/singer, even if you don’t like her?
Is there something missing from this data-driven argument?
Thank you for reading this article! I hope you found this to be interesting. Any comments or critiques are warmly welcomed.
Additionally, feel free to approach me for a chat on any of these topics, and to connect with me on LinkedIn .
