Beatles lyrics

purrr_geniusR.R

library(tidyverse)
library(scales)
theme_set(silgelib::theme_plex())
library(geniusR)
library(tidytext)

albums <-  tibble(artist = rep("The Beatles", 13),
                  album = c("Please Please Me",
                            "With the Beatles",
                            "A Hard Day's Night",
                            "Beatles for Sale",
                            "Help!", 
                            "Rubber Soul",
                            "Revolver",
                            "Sgt. Pepper's Lonely Hearts Club Band",
                            "Magical Mystery Tour",
                            "The Beatles (The White Album)",
                            "Yellow Submarine",
                            "Abbey Road",
                            "Let It Be"),
                  year = c(1963,
                           1963,
                           1964,
                           1964,
                           1965,
                           1965,
                           1966,
                           1967,
                           1967,
                           1968,
                           1969,
                           1969,
                           1970))

albums

album_lyrics <- albums %>% 
    mutate(tracks = map2(artist, album, genius_album))

album_lyrics

tidy_lyrics <- album_lyrics %>%
    unnest(tracks) %>%
    unnest_tokens(word, lyric)


words_by_year <- tidy_lyrics %>%
    count(year, word) %>%
    anti_join(get_stopwords()) %>%
    complete(year, word, fill = list(n = 0)) %>%
    group_by(word) %>%
    mutate(word_total = sum(n)) %>%
    ungroup %>%
    filter(word_total > 70) %>%
    group_by(year) %>%
    mutate(year_total = sum(n)) %>%
    ungroup %>%
    rename(count = n)

library(broom)

nested_models <- words_by_year %>%
    nest(-word) %>%
    mutate(models = map(data, ~ glm(cbind(count, year_total) ~ year, ., 
                                    family = "binomial")))

slopes <- nested_models %>%
    unnest(map(models, tidy)) %>%
    filter(term == "year") %>%
    arrange(estimate) %>%
    filter(p.value < 0.05)
    
    library(ggrepel)

plot_words <- slopes %>%
    group_by(sign = estimate > 0) %>%
    top_n(5, abs(estimate)) %>%
    ungroup %>%
    mutate(sign = ifelse(sign, "Increasing", "Decreasing")) %>%
    inner_join(words_by_year) %>%
    mutate(freq = count / year_total)

plot_labels <- plot_words %>% 
    group_by(word) %>% 
    mutate(yearselect = ifelse(sign == "Decreasing", 
                         min(year), 
                         max(year))) %>% 
    group_by(word) %>% 
    filter(year == yearselect) %>%
    ungroup

ggplot(plot_words, aes(year, freq,
                       color = sign, group = word)) +
    geom_smooth(aes(year, freq, group = sign, color = sign), 
                method = "lm", lty = 2, alpha = 0.2, size = 0.8) +
    geom_line(alpha = 0.8, size = 1.3, show.legend = FALSE) +
    geom_text_repel(data = filter(plot_labels, sign == "Decreasing"),
                    aes(year, freq, label = word, color = sign),
                    nudge_x = -0.5,
                    hjust = 0,
                    family = "IBMPlexSans") +
    geom_text_repel(data = filter(plot_labels, sign == "Increasing"),
                    aes(year, freq, label = word, color = sign),
                    nudge_x = 0.5,
                    hjust = 1,
                    family = "IBMPlexSans") +
    facet_wrap(~ sign, nrow = 2, scales = "free_y") +
    scale_y_continuous(labels = scales::percent_format()) +
    theme(legend.position="none") +
    labs(x = NULL,
         y = "Word frequency",
         title = "Words changing in frequency in Beatles' lyrics",
         subtitle = "The Beatles sang less about babies and more about getting back somewhere as time passed")