TidyTuesday Wk52: Christmas Songs

Get the Data

christmas_songs <- readr::read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-12-24/christmas_songs.csv")

## Parsed with column specification:
## cols(
##   url = col_character(),
##   weekid = col_character(),
##   week_position = col_double(),
##   song = col_character(),
##   performer = col_character(),
##   songid = col_character(),
##   instance = col_double(),
##   previous_week_position = col_double(),
##   peak_position = col_double(),
##   weeks_on_chart = col_double(),
##   year = col_double(),
##   month = col_double(),
##   day = col_double()
## )

christmas_lyrics <- readr::read_tsv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-12-24/christmas_lyrics.tsv")

## Parsed with column specification:
## cols(
##   title = col_character(),
##   artist = col_character(),
##   songid = col_character(),
##   weekid = col_character(),
##   track_title = col_character(),
##   track_n = col_double(),
##   line = col_double(),
##   lyric = col_character()
## )

library(ggplot2)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

labs = tibble(
  x = c(2001.5,2005.5,2012.5,2008.5), y = c(33,3.5,7,33),
  label = c('My sissy was the reason\n for this billboard hit ',
            'My #2 holiday song',
            'Christmas songs seem to\nbe ranking in the top 50\nmore often the past 8 years\n(more red dots)',
            '2011 was super festive!\nNINE Christmas Songs\n making the top 100')
)


christmas_songs %>%
  filter(year >= 1989)%>%
  ggplot( aes(year,song))+
    geom_point(aes(color = week_position), size = 6)+
    labs(title = "HIT CHRISTMAS SONGS SINCE I WAS BORN", y = "",x = "Year the Song Hit Billboard's Top 100")+
    scale_color_gradientn(colours = c("darkred","red","darkgreen", "lightgrey"),name = "Billboard\nRank")+
    scale_x_continuous(breaks=c(1990, 1995, 2000, 2005, 2010, 2015))+
    geom_curve(x = 2004, y = 33, xend = 2001.5, yend = 38, arrow = arrow(length = unit (.2,'cm')), curvature = 0.5)+
    geom_curve(x = 2009, y = 3.5, xend = 2010, yend = 4.1, arrow = arrow(length = unit (.2,'cm')), curvature = 0.3)+
    geom_curve(x = 2009.5, y = 31, xend = 2010.5, yend = 24, arrow = arrow(length = unit (.2,'cm')), curvature = 0.3)+
  geom_label(data = labs, aes(label = label, x = x, y = y), hjust = 0, size = 3, label.size = 0)+
  theme_bw()+
  theme(title = element_text(colour = "darkred"),
        axis.text = element_text(colour = c("darkred","darkgreen")),
        panel.border = element_rect(colour = "darkgreen"))

ggsave("20191224_TidyTuesdayWk52_christmassongs.PNG", width = 14, height = 6, units = "in", dpi = 300)