Pop Isn’t as Formulaic as People Believe — And The Data Proves It
I won’t lie. I’ve often been guilty of shaming pop for being too unoriginal and, as a result, boring. I mean, every week, the top songs seem to use the same drum beat and chords. The lyrical themes rarely differ, and every song seems to make similar stylistic choices. I do, however, respect everyone’s music taste, which is why I’m trying to be more aware of when I make such statements to my friends and family. Still, though, the idea lies buried in the back of my mind — I can’t get rid of it. So, for my own peace of mind, I decided to finally answer the question once and for all: is pop really more formulaic than other music I like? For this analysis, I chose to compare pop with dream pop, which is arguably my favorite genre of all.
Note #1: I know pop has quite a vague definition, but I am sticking to Spotify’s classification of songs.
Note #2: If you would like to skip the technical explanation and go straight to the results, scroll until you see a graph.
To carry out my analysis, I chose to use this dataset, which provides a wide range of information collected from Spotify. Specifically, the values of interest to me were those that could tell me anything about the musical qualities of a song. In this case, these qualities were valence (how happy a song is), acoustics, danceability, energy, and instrumental-ness. But obtaining these values wasn’t going to tell me very much; I still couldn’t answer my question by merely collecting tons of these values from different songs. So, I decided that, for each genre (pop and dream pop), I would take the standard deviation for each of these values.
For those who may not be familiar with standard deviation or need a refresher, standard deviation basically tells you how spread out, or varied, the data is. In our case, a higher standard deviation essentially shows higher originality within the genre; if it turns out that the standard deviations of these values are significantly lower for pop music, we can conclude that pop music really isn’t as innovative as dream pop since most songs in the genre don’t differ as much as dream-pop songs.
Because the dataset did not provide information about what genre(s) each song corresponded to, I had to come up with an alternate way to get that information. I did this the same way as I did in my previous analysis: I identified all artists labeled as pop artists and then did the same for dream pop artists using artists.csv. Once again, I only selected artists with a Spotify popularity score greater than 10 so that I’d only have somewhat relevant artists.
Now that I had two lists, one of all the pop artists and the other of dream pop artists, I turned to tracks.csv. At this point, I had to decide how many songs I wanted to look at. I chose to go with 1000 of the most popular songs from each genre since I figured that going with too many songs would make for an unfair comparison as, due to the reduced popularity of dream pop, I would end up choosing much less relevant dream-pop songs than pop songs. At this point, I had two more lists, each containing the IDs of the top 1000 songs from each genre.
After that, I had to switch to data_o.csv, as track.csv did not contain enough information about each song’s musical qualities. So, using my lists of song IDs, I cross-referenced each song in data_o.csv and retrieved information about valence, acoustics, danceability, energy, and instrumental-ness.
At this point, I had values for all the qualities for each song I was considering, and so I had all the information I would need. For each genre, I then calculated the standard deviation for each of these qualities. Great, I now have precisely what I sought out at the beginning. Here are the findings.
Note: All the qualities are measured by Spotify on a scale of 0.0–1.0. The standard deviation value follows that same scale.
At surface level, the graph seems to show that dream-pop songs, on average, are more willing to step outside the box. This isn’t so surprising, given that many dream pop bands are/were signed to smaller record labels when compared to the pop giants. Beach House, for example, who is arguably the most influential active dream-pop band (and also my favorite band of all time), is signed to SubPop. And although SubPop is definitely not a small label by any means, record labels like Warner, Sony, and Universal inarguably dwarf it.
With smaller labels, artists are often gifted more creative freedom — at least more than artists signed to major labels who work to ensure that every song they release makes as much money as possible. This isn’t to say this is always true, but it is for the most part.
However, I can’t just make such a conclusion based on how the graph looks. After all, most of these values aren’t that different between the two genres. So, I used the F-test, a statistical test that allows for the comparison of standard deviations to see if there is a statistically significant difference between them. I won’t go into any detail about how the test works since that’s not really important, but if you’re interested, you can read more about it here.
The test concluded that for all qualities but instrumental-ness, there is not a statistically significant difference between the standard deviations. What does this mean? Well, it means that pop takes no fewer risks than dream-pop for five out of six musical qualities. Let’s look at each of these qualities in detail to see why these results may be the way they are.
The first quality is valence. A lot of pop is definitely upbeat, like “Body,” but no small number of pop songs are also sad ballads (“Someone You Loved”). The converse goes for dream pop, which, for the most part, can be slightly moodier (think Slowdive, Mazzy Star, for example). Nevertheless, a lot of dream pop isn’t. One of my favorites, The Sundays, wrote some pretty upbeat dream pop in their debut album, Reading, Writing and Arithmetic — and they were certainly not the only ones. Still, we see that both genres can vary quite a bit in terms of how positive a song feels. It thus makes sense that these two genres have similar standard deviations for valence.
Next, we have acoustics. I’m actually slightly surprised that both these genres have decently high standard deviations for this quality. I think a lot of pop nowadays relies a lot on synths, heavy production, and so on. The same goes for dream pop, which is often built on electric guitars, synths, and other non-acoustic sounds. I certainly can’t think of too many dream pop acts who have heavily used more acoustic sounds to justify such a standard deviation. If anyone knows any, please do let me know so I can check them out.
Third, we have danceability. Both these genres seem to be pretty consistent with their danceability. Using the data from intermediate steps in my analysis, we can see that pop generally scores higher on danceability while dream pop scores lower. Again, this isn’t news. You simply just can’t dance as hard to the Cocteau Twins as you can Dua Lipa. It’s unfortunate, but it’s true.
Energy is the last quality where pop and dream pop don’t show statistically significant differences in their standard deviation. Spotify defines energetic songs as songs that “feel fast, loud, and noisy.” A lot of pop certainly fits that bill. But again, songs like Billie Eilish’s “when the party’s over” certainly do not. With dream pop, you have loud, fast, and pulsating songs like “10 Mile Stereo,” but you also have Julee Cruise’s “Falling” on the other end of the spectrum. So, it makes that both these genres have decently high standard deviations for energy.
Finally, the one difference: instrumental-ness. The data clearly shows that dream-pop songs are far more varied in terms of instrumental-ness, which Spotify defines to be the amount of vocals in a song (the fewer vocals, the higher the instrumental-ness). Pop, for the most part, is highly centered around vocals. You’ll rarely if ever, find a song with considerable periods without vocals. This may be attributed to several factors, one of which may be that major labels often don’t allow their artists to do this as it can lower the chances of the song doing well in the mainstream (and thus potentially decrease profits).
Dream pop, on the other hand, is much more focused on atmosphere, and that can be achieved with or without vocals. Songs from Galaxie 500’s On Fire, for example, often play out as a dialogue between the instruments and the vocalist, Dean Wareham. Neither one nor the other takes center stage, and that’s part of what makes the album work so well. On the other hand, the Cocteau Twins often (not always) have vocals that run throughout for pretty much the whole song. The majority of Heaven or Las Vegas certainly demonstrates this. It is thus clear, then, why dream-pop shows so much more diversity in their levels of instrumental-ness.
Regardless, the fact that dream-pop showed no more originality than pop in five out of six of these qualities came as a shock to me, given how much I adore dream pop. I thought to myself, “how can a genre that is literally my favorite ever be no barely any more innovative than a genre I’ve believed to be so boring and formulaic?” Well, the answer might just be that pop isn’t as uninspired as I and many others believe it to be. I think I wanted to believe so bad that pop was boring that I eventually did; I certainly won’t be doing that anymore. This was a big wake-up call for me, and I hope it will be for others who shame pop too.
If anyone is interested in seeing how their favorite genre stacks up against pop (or any genre for that matter), let me know in the comments and I can try to do something for you!
