An Analysis of Modern Music Criticism

Appendix to report

Data cleaning

Fantano data set:

library(tidyverse)
library(skimr)
library(lubridate)

fantano_raw1 <- fantano_raw |> 
  # Extract data and convert to correct format
  separate(date, into = c("date", "time"), sep = " ") |>
  mutate(date = as.Date(date, format = "%Y-%m-%d")) |>
  # convert score to numeric
  mutate(score = parse_number(score)) |>
  mutate(score = ifelse(score > 10, 10, score)) |>
  # Keep the six most important tags
  separate(tags, into = c("tag1", "tag2", "tag3", "tag4", "tag5", "tag6"), sep = "\\|") |>
  select(-time, -url) |>
  # Delete rows with NAs
  na.omit() |>
  arrange(desc(score))

num_in_each_genre <- fantano_raw1 |>
  group_by(genre) |>
  summarize(n = n())

fantano_raw2 <- fantano_raw1 |> 
  left_join(num_in_each_genre, by = "genre")

# Merge genre with less than 10 songs with other genres
fantano_raw2 <- fantano_raw2 |>
  mutate(genre = ifelse(n < 10 & grepl("electronic", genre), "electronic", genre)) |>
  mutate(genre = ifelse(n < 10 & grepl("hip hop", genre), "hip hop", genre)) |>
  mutate(genre = ifelse(n < 10 & grepl("loud rock", genre), "loud rock", genre)) |>
  mutate(genre = ifelse(n < 10 & grepl("pop", genre), "pop", genre)) |>
  mutate(genre = ifelse(n < 10 & grepl("rock", genre), "rock", genre)) |>
  mutate(genre = ifelse(n < 10 & grepl("sub pop", genre), "sub pop", genre)) |>
  mutate(genre = ifelse(n < 10 & !grepl("electronic", genre) & !grepl("hip hop", genre) & !grepl("loud rock", genre) & !grepl("pop", genre) & !grepl("rock", genre) & !grepl("sub pop", genre), "other", genre))

# Put other at last since it does not provide useful information
fantano_raw2$genre <- factor(fantano_raw2$genre, levels = c("electronic", "hip hop", "loud rock", "pop", "rock", "sub pop", "other"))

fantano_cleaned <- fantano_raw2 |>
  select(-n, -name)

Pitchfork data set:

# Delete repetitive labels
pitchfork_raw1 <- pitchfork_raw|>
  separate_rows(label, sep = ",") |>
  group_by(album) |>
  summarise(label = paste(unique(label), collapse = ","))

pitchfork_raw2 <- pitchfork_raw|>
  select(-link, - review)|>
  # Extract date and convert to correct format
  separate(date, into = c("review_month", "review_day","review_year"), sep = " ")|>  
  mutate(
         review_month = case_when(
         review_month == "January" ~ "01",
         review_month == "February" ~ "02",
         review_month == "March" ~ "03",
         review_month == "April" ~ "04",
         review_month == "May" ~ "05",
         review_month == "June" ~ "06",
         review_month == "July" ~ "07",
         review_month == "August" ~ "08",
         review_month == "September" ~ "09",
         review_month == "October" ~ "10",
         review_month == "November" ~ "11",
         review_month == "December" ~ "12"),
         review_day = as.numeric(review_day),
         review_month = as.numeric(review_month),
         review_year = as.numeric(review_year),
         date = as.Date(with(pitchfork_raw, paste(review_year, review_month,
                                                  review_day,sep="-")), "%Y-%m-%d"),
         # Compare the release and review year
         is_retrospective = if_else(release_year == review_year, "no", "yes"))|>
         # Exclude irrelevant variables
         select(-review_month, -review_day,-label)
    
# Joined pitchfork_raw1 and pitchfork_raw2 by album
pitchfork_cleaned <- inner_join(x = pitchfork_raw1, y = pitchfork_raw2, 
                       by = c("album"))|>
  # Delete rows with NAs
      na.omit() |>
      arrange(desc(score))|>
  # Keep the six most important tags
    separate(label, into = c("tag1", "tag2", "tag3", "tag4", "tag5", "tag6"), sep = "\\,")

# export the cleaned datasets
write_csv(fantano_cleaned, file = "data/fantano-cleaned.csv")
write_csv(pitchfork_cleaned, file = "data/pitchfork-cleaned.csv")

Other appendices (as necessary)