Top Songs of the 1950’s

Top 5 Songs

Decade Rank	Title	Artist	Genre	Year
1	Put Your Head On My Shoulder	Paul Anka	adult standards	2000
2	Whatever Will Be Will Be (Que Sera Sera) (with Frank DeVol & His Orchestra) - Single Version	Doris Day	adult standards	1948
3	Everybody Loves Somebody	Dean Martin	adult standards	2013
4	Take Good Care Of My Baby - 1990 Remastered	Bobby Vee	adult standards	2011
5	A Teenager In Love	Dion & The Belmonts	adult standards	1959

Note: If year of song is outside of 1950, then year corresponds to re-release year (not original release year).

Most Popular Genres

Here are the genre types that make up the top songs of the decade:

Spotify specifies that ‘adult standards’, which is a genre defined by classic or oldschool hits of the past, was popular in the 1950’s. However, this genre is aggregating the songs of the 1950’s (along with ‘old’ music of other decades) into a contemporary genre called ‘adult standards’, as they are ‘throwback’ songs relative to now. It is better to treat this genre as ‘unspecified’, as Spotify has not provided the genre of this music in the context of the 1950’s.

While the songs don’t cluster into one contextually specific genre, it is interesting that some genres appear in the 1950’s, like doo-wop, merseybeat and classic soul, that become relics of the past - they phase out of popularity in the coming decades.

What Makes a Song Popular

Let’s unpack what actually made a song popular in the 1950’s. What elements of the music were important for it to be liked by the masses throughout a whole decade? To explore this, I ran a stepwise regression on all the audio features Spotify provided for the Top 100 songs of the 1950’s, with popularity as the dependent variable.

After running the analysis, we are left with the following co-efficients for a regression formula:

\[ popularity = 26.60 + (-0.10)(duration) \] This equation lets us make the inference that in the 1950’s, the audio features of a track that contributed to it being popular were duration only. More specifically, a song being shorter in duration would make it more popular.

A summary of the stepwise analysis is presented below:

## 
## Call:
## lm(formula = popularity ~ duration, data = spotify_data_clean %>% 
##     filter(decade == decade_input))
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -16.338  -7.812  -1.477   6.330  30.616 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept) 26.59964    8.07229   3.295  0.00154 **
## duration     0.09539    0.05339   1.787  0.07827 . 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 10.2 on 71 degrees of freedom
## Multiple R-squared:  0.04302,    Adjusted R-squared:  0.02954 
## F-statistic: 3.192 on 1 and 71 DF,  p-value: 0.07827