Exploring the World of Music

Introduce the topic and motivation

• We all have a passion for music!

• Research question: do the duration of a song, the tempo of a song, artist familiarity, and artist popularity have an association with the popularity of a song?

Introduce the data

• The dataset was created to analyze the trend of music listeners and to encourage research on algorithms that scale to commercial sizes and to derive data points from 3,064 songs.

• We collected data from Million Song Data Set which is funded by Echo Nest. It is possible that a great portion of our data is only from Spotify.

• The observations are songs and the attributes provide information about songs (song.id, song.year, song.tempo, song.hotttnesss) and their correlating artists (artist.id, artist.hotttnesss, artist.familiarity)

Highlights from EDA

── Attaching packages ─────────────────────────────────────── tidyverse 1.3.2 ──
✔ ggplot2 3.4.0     ✔ purrr   1.0.1
✔ tibble  3.2.1     ✔ dplyr   1.1.2
✔ tidyr   1.3.0     ✔ stringr 1.5.0
✔ readr   2.1.3     ✔ forcats 0.5.2
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
── Attaching packages ────────────────────────────────────── tidymodels 1.0.0 ──

✔ broom        1.0.2     ✔ rsample      1.1.1
✔ dials        1.1.0     ✔ tune         1.1.1
✔ infer        1.0.4     ✔ workflows    1.1.2
✔ modeldata    1.0.1     ✔ workflowsets 1.0.0
✔ parsnip      1.1.0     ✔ yardstick    1.1.0
✔ recipes      1.0.6     

── Conflicts ───────────────────────────────────────── tidymodels_conflicts() ──
✖ scales::discard() masks purrr::discard()
✖ dplyr::filter()   masks stats::filter()
✖ recipes::fixed()  masks stringr::fixed()
✖ dplyr::lag()      masks stats::lag()
✖ yardstick::spec() masks readr::spec()
✖ recipes::step()   masks stats::step()
• Use suppressPackageStartupMessages() to eliminate package startup messages

Loading required package: airports

Loading required package: cherryblossom

Loading required package: usdata


Attaching package: 'openintro'


The following object is masked from 'package:modeldata':

    ames


New names:
Rows: 2693 Columns: 12
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (5): artist.name, artist.id, artist.terms, song.id, genre
dbl (7): ...1, artist.familiarity, artist.hotttnesss, song.hotttnesss, song....

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
`geom_smooth()` using method = 'gam' and formula = 'y ~ s(x, bs = "cs")'

`geom_smooth()` using formula = 'y ~ x'

`geom_smooth()` using formula = 'y ~ x'

`geom_smooth()` using formula = 'y ~ x'

Hypothesis Testing

• We want to examine the relationship between song tempo and popularity

• Test for significant difference between the proportion of songs with hotness greater than 0.8 inside and outside 100-140 bpm tempo.

• We chose this range because this is the average tempo range of a song, and we think that 0.8 is the threshold for popularity.

• p-value is 0.136. The data does not provide evidence of a significant difference in popularity of songs with tempo within and outside 100-140 bpm.

Hypothesis Testing

# A tibble: 1 × 1
  p_value
    <dbl>
1   0.132

# A tibble: 1 × 2
  lower_ci upper_ci
     <dbl>    <dbl>
1 -0.00160   0.0256

Conclusions + future work

Conclusions

• The hottest songs are usually 240 seconds long

• Song tempo and duration don’t have a significant effect on song popularity

• Artist popularity and familiarity do have a relationship with song popularity

• Artists and music labels should focus on building their popularity and retention listening rates to increase their song popularity.

Future Work