italianRapSong.Rmd

---
title: "Spotify italian rap"
author: "Giovanni Esposito"
output:
  revealjs::revealjs_presentation:
    theme: moon
    highlight: haddock
editor_options:
  chunk_output_type: inline
---

```{r setup, include=FALSE}
# cache results
knitr::opts_chunk$set(
  cache = TRUE,
  echo = TRUE,
  eval = TRUE,
  message = FALSE,
  warning = FALSE,
  fig.align = 'center'
)

```


```{r library, include=FALSE}
library(tidyverse)
library(ggplot2)
library(plotly)
library(gapminder)
library(gridExtra)
library(grid)
library(reshape2)
library(RColorBrewer)
library(scales)

```
    
## Dataset

<h3>
Dataset was created use https://github.com/gioggi/SpotifyScarper.<br/>
Dataset is update to april 2021. <br/>
about 160 Artists<br/>
about 11.5k Tracks
</h3>
<hr/>
<h4>
<b>Key helper</b>:<br/>
<b>Popularity</b>: Measure the popularity based on play number of the track <br/>
<b>Valence</b>: a measure of happiness  <br/>
<b>Liveness</b>: detects the presence of an audience in the recording.  <br/>
<b>Speechiness</b>: Detects the presence of spoken words in a track. <br/>
<b>Energy</b>: represents a perceptual measure of intensity and activity. <br/>
<b>Danceability</b>: describes how suitable a track is for dancing based <br/>
</h4>

```{r dataset, message=FALSE, echo=FALSE}


# put artist and into a data frame and columns mapping
artists = read_csv("/home/gioggi/Documents/ItalianRapSong/data/artist210421.csv")
artists_mapped = rename(
  artists,
  "artist_id" = "Id",
  "artist_name" = "Name",
  "artist_popularity" = "Popularity",
  "artist_followers" = "Followers",
  "artist_geners" = "Generes"
)
artist_color <- rainbow(nrow(artists_mapped))
artists_mapped = cbind(artists_mapped, artist_color)
# put track and into a data frame
tracks = read_csv("/home/gioggi/Documents/ItalianRapSong/data/track210421.csv") %>%
  group_by(Id)

tracks$album_release_date <-
  as.Date(tracks$album_release_date, '%Y')
# left join between tracks and artists
# merge( x= tracks, y = artists_mapped, by = "artist_id", all.x = TRUE, fill=-9999)
```


## How are the 50 more popular italian rap artist now for Spotify?

```{r artist by popularity, message=FALSE, echo=FALSE}
# order artist by popularity
popular_artists = head(artists_mapped[order(artists_mapped$artist_popularity, decreasing = T),], n = 50)

# chart
ggplot(data = popular_artists, aes(x = artist_popularity, y = reorder(artist_name, artist_popularity))) +
  geom_bar(stat = "identity", fill = popular_artists$artist_color) +
  theme_dark() +
  theme(
    axis.line = element_blank(),
    axis.title.y = element_blank()
  ) +
  xlab('Artist popularity')
```


## How are the 50 more followed italian rap artist now for Spotify?

```{r artist by followers, message=FALSE, echo=FALSE}
# order artist by followers
followest_artists = head(artists_mapped[order(artists_mapped$artist_followers, decreasing = T), ], n = 50)

# chart
ggplot(data = followest_artists, aes(x = artist_followers, y = reorder(artist_name, artist_followers))) +
  geom_bar(stat = "identity", fill = followest_artists$artist_color) +
  theme_dark()  +
  theme(
    axis.line = element_blank(),
    axis.title.y = element_blank()
  )+
  xlab('Number of followers')

```

# Matched popularity and followers for top 50 artists!
```{css, echo=FALSE}
    .slides {
      zoom: 1 !important;
    }

```
```{r match popularity and followers, message=FALSE, echo=FALSE}
artist_songs_number = merge(
  x = tracks,
  y = popular_artists,
  by = "artist_id",
  all.x = FALSE,
  fill = -9999
) %>%
  count(artist_id, name = "n_songs")
artist_sum_popularity = merge(
  x = tracks,
  y = popular_artists,
  by = "artist_id",
  all.x = FALSE,
  fill = -9999
) %>%
  group_by(artist_id) %>%
  summarise(tot_popularity = sum(popularity))

top_artists = merge(
  x = artist_songs_number,
  y = popular_artists,
  by = "artist_id",
  all.x = FALSE,
  fill = -9999
)
top_artists = merge(
  x = artist_sum_popularity,
  y = top_artists,
  by = "artist_id",
  all.x = FALSE,
  fill = -9999
)
p <- top_artists %>%
ggplot(aes(tot_popularity, artist_followers, size = n_songs, color=artist_color, label=artist_name),x.title="aw") +
  geom_point() +
  theme_dark() +
  theme(legend.position='none') +
  xlab('Total popularity')
ggplotly(p,tooltip = c("artist_name","n_songs","artist_followers","tot_popularity"))

```


## Matched average popularity and followers for top 50 artists!
```{r match popularity and followers with average, message=FALSE, echo=FALSE}
top_artists['average_popularity'] = top_artists['tot_popularity'] / top_artists['n_songs']
p <- top_artists %>%
ggplot(aes(average_popularity, artist_followers, size = n_songs, color=artist_color, label=artist_name),x.title="aw") +
  geom_point() +
  theme_dark() +
  theme(legend.position='none') +
  xlab('Average popularity')
ggplotly(p,tooltip = c("artist_name","n_songs","artist_followers","tot_popularity"))

```


# The stats for all songs by Spotify!
```{r point stats, message=FALSE, echo=FALSE}

danceability <-
  ggplot(tracks, aes(x = popularity, y = danceability)) +
  geom_point(color = "purple", alpha = 0.2) +
  geom_smooth(se = FALSE) +
  theme_dark()+
  xlab('Popularity')+
  ylab('Danceability')
energy <- ggplot(tracks, aes(x = popularity, y = energy)) +
  geom_point(color = "red", alpha = 0.2) +
  geom_smooth(se = FALSE) +
  theme_dark()  +
  xlab('Popularity')+
  ylab('Energy')
speechiness <-
  ggplot(tracks, aes(x = popularity, y = speechiness)) +
  geom_point(color = "lightblue", alpha = 0.2) +
  geom_smooth(se = FALSE) +
  theme_dark()+
  xlab('Popularity')+
  ylab('Speechiness')
liveness <- ggplot(tracks, aes(x = popularity, y = liveness)) +
  geom_point(color = "green", alpha = 0.2) +
  geom_smooth(se = FALSE) +
  theme_dark()+
  xlab('Popularity')+
  ylab('Liveness')
valence <- ggplot(tracks, aes(x = popularity, y = valence)) +
  geom_point(color = "blue", alpha = 0.2) +
  geom_smooth(se = FALSE) +
  theme_dark()+
  xlab('Popularity')+
  ylab('Valence')
duration_ms <-
  ggplot(tracks, aes(x = popularity, y = duration_ms)) +
  geom_point(color = "orange", alpha = 0.2) +
  geom_smooth(se = FALSE) +
  theme_dark()+
  xlab('Popularity')+
  ylab('Duration ms')

grid.arrange(danceability,
             energy,
             speechiness,
             liveness,
             valence,
             duration_ms)

```



## Spotify stats for top 200 popular tracks!
```{r boxplot stats top 200 populars, message=FALSE, echo=FALSE}


top_200_popular_tracks = head(tracks[order(tracks$popularity, decreasing = T), ], n = 200)


top_danceability <-
  ggplot(top_200_popular_tracks, aes(x = popularity, y = danceability)) +
  geom_boxplot(fill = "purple") +
  theme_dark() +
  xlab('Popularity')+
  ylab('Danceability')
top_energy <-
  ggplot(top_200_popular_tracks, aes(x = popularity, y = energy)) +
  geom_boxplot(fill = "red") +
  theme_dark()  +
  xlab('Popularity')+
  ylab('Energy')
top_speechiness <-
  ggplot(top_200_popular_tracks, aes(x = popularity, y = speechiness)) +
  geom_boxplot(fill = "lightblue") +
  theme_dark() +
  xlab('Popularity')+
  ylab('Speechiness')
top_liveness <-
  ggplot(top_200_popular_tracks, aes(x = popularity, y = liveness)) +
  geom_boxplot(fill = "green") +
  theme_dark() +
  xlab('Popularity')+
  ylab('Liveness')
top_valence <-
  ggplot(top_200_popular_tracks, aes(x = popularity, y = valence)) +
  geom_boxplot(fill = "blue") +
  theme_dark() +
  xlab('Popularity')+
  ylab('Valence')
top_duration_ms <-
  ggplot(top_200_popular_tracks, aes(x = popularity, y = duration_ms)) +
  geom_boxplot(fill = "orange") +
  theme_dark()+
  xlab('Popularity')+
  ylab('Duration ms')

grid.arrange(
  top_danceability,
  top_energy,
  top_speechiness,
  top_liveness,
  top_valence,
  top_duration_ms
)

```

# Spotify median stats in the time!
```{r stats by date, message=FALSE, echo=FALSE}


year_median_danceability = tracks %>%
  group_by(album_release_date) %>%
  summarise(median_danceability = mean(danceability))

year_median_danceability_charts <-
  ggplot(year_median_danceability,
         aes( album_release_date, median_danceability)) +
  geom_line(color = "purple", size = 1.2) +
  theme_dark()+
  xlab('Album release date')+
  ylab('Danceability')

year_median_energy = tracks %>%
  group_by(album_release_date) %>%
  summarise(median_energy = mean(energy))

year_median_energy_charts <-
  ggplot(year_median_energy, aes( album_release_date, median_energy)) +
  geom_line(color = "red", size = 1.2) +
  theme_dark()+
  xlab('Album release date')+
  ylab('Energy')

year_median_speechiness = tracks %>%
  group_by(album_release_date) %>%
  summarise(median_speechiness = mean(speechiness))

year_median_speechiness_charts <-
  ggplot(year_median_speechiness,
         aes(album_release_date,median_speechiness)) +
  geom_line(color = "lightblue", size = 1.2) +
  theme_dark()+
  xlab('Album release date')+
  ylab('Speechiness')

year_median_liveness = tracks %>%
  group_by(album_release_date) %>%
  summarise(median_liveness = mean(liveness))

year_median_liveness_charts <-
  ggplot(year_median_liveness, aes(album_release_date, median_liveness)) +
  geom_line(color = "green", size = 1.2) +
  theme_dark()+
  xlab('Album release date')+
  ylab('Liveness')

year_median_valence = tracks %>%
  group_by(album_release_date) %>%
  summarise(median_valence = mean(valence))

year_median_valence_charts <-
  ggplot(year_median_valence, aes(album_release_date,median_valence)) +
  geom_line(color = "blue", size = 1.2) +
  theme_dark()+
  xlab('Album release date')+
  ylab('Valence')

year_median_duration_ms = tracks %>%
  group_by(album_release_date) %>%
  summarise(median_duration_ms = mean(duration_ms))

year_median_duration_ms_charts <-
  ggplot(year_median_duration_ms,
         aes(album_release_date, median_duration_ms)) +
  geom_line(color = "orange", size = 1.2) +
  theme_dark()+
  xlab('Album release date')+
  ylab('Duration ms')

grid.arrange(
  year_median_danceability_charts,
  year_median_energy_charts,
  year_median_speechiness_charts,
  year_median_liveness_charts,
  year_median_valence_charts,
  year_median_duration_ms_charts
)

```