Introduction

When hearing the songs at the 2022 Brit awards I was struck by the dullness of it all. This was not a result of the change in the award structure. The new format does appear quite reasonable, as there is no reason to separate artists on gender. However the lack of energy and sheer fun in the performances hit me. It all was very dull and lacking in joy. Of course, this may be a subjective judgement based on my own personal taste.

Spotify uses algorithms to classify songs according to certain audio features, including energy, valence (cheeriness) and danceability. I therefore could use non subjective scores to evaluate my gut feeling.

I downloaded a, somewhat cherry picked, but interesting selection of artists including the main winners of the 2022 awards (Adele, Wolf Alice, Little Syms) and some notable previous winners. Radiohead was added as a reference, as the band are generally regarded to produce the most depressing material of all time.

With regard to energy, my gut feeling appears to be well supported by the data. The 2022 winners produced the least energetic music of all the selected artists, scoring even worse than Radiohead. Radiohead, as would be expected, scored the lowest in terms of cheerfulness (valence). Ed Sheeran obtained a surprising second place with regard to danceability, although whether dad dancing counts as dancing is debatable. Overall, I think that the data supports my gut feeling that popular music is becoming less energetic and generally more depressing.

d1 <- data.frame(artist="Oasis",get_artist_audio_features('Oasis'))
d2 <- data.frame(artist="Blur",get_artist_audio_features('blur'))
d3 <- data.frame(artist="Arctic Monkeys",get_artist_audio_features('Arctic Monkeys'))
d4 <- data.frame(artist="Adele",get_artist_audio_features('Adele'))
d5 <- data.frame(artist="Ed Sheeran",get_artist_audio_features('Ed Sheeran'))
d6 <- data.frame(artist="Little Simz",get_artist_audio_features('Little Simz'))
d7 <- data.frame(artist="Little Mix",get_artist_audio_features('Little Mix'))
d8 <- data.frame(artist="Radiohead",get_artist_audio_features('Radiohead'))
d9 <- data.frame(artist="Wolf Alice",get_artist_audio_features('Wolf Alice'))
d10 <- data.frame(artist="Spice Girls",get_artist_audio_features('Spice Girls'))

d<-rbind(d1,d2,d3,d4,d5,d6,d7,d8,d9, d10)
d$track_name<-gsub("Remastered","",d$track_name)
d$album_name<-gsub("(Remastered)","",d$album_name)
library(dplyr)
d<-d[,c(1,3,8,10:21,31)]

Results

Energy

Boxplots

library(forcats)
library(ggforce)
d$artist= fct_reorder(d$artist,d$energy,median,TRUE)
d %>% ggplot(aes(x=energy, y=artist)) ->g0
g0+geom_boxplot()

Confidence intervals for the mean

ci(g0)

Valence

Boxplots

library(forcats)
library(ggforce)
d$artist= fct_reorder(d$artist,d$valence,median,TRUE)
d %>% ggplot(aes(x=valence, y=artist)) ->g0
g0+geom_boxplot()

Confidence intervals for the mean

ci(g0)

Danceability

Boxplots

library(forcats)
library(ggforce)
d$artist= fct_reorder(d$artist,d$danceability,median,TRUE)
d %>% ggplot(aes(x=danceability, y=artist)) ->g0
g0+geom_boxplot()

Confidence intervals for the mean

ci(g0)