Rapid Decline in Popularity of Names

name_falls.tsv

name_popularity_fall.R

pacman::p_load("tidyverse","fs")

national_files = dir_ls(path(data_dir, "national"), regex = "yob\\d{4}\\.txt")

national_spec = cols(name = col_character(),
                     gender = col_factor(),
                     n = col_integer())

national = read_csv(
    national_files,
    col_names = names(national_spec$cols),
    col_types = national_spec,
    id = "file_name"
) |>
    mutate(
        yr = str_extract(file_name, "\\d{4}") |> as.integer(),
        gender = fct_recode(gender, male = "M", female = "F"),
        .keep = "unused"
    ) |>
    select(yr, gender, name, n)


national_w_rank = national |> 
    mutate(rank = dense_rank(desc(n)), .by=c(yr,gender)) |> 
    select(-n)

national_peak_yr = national_w_rank |> 
    arrange(name, gender, yr) |>
    filter(rank == min(rank), .by=c(name, gender)) |> 
    filter(row_number() == max(row_number()), .by=c(name, gender)) |> 
    select(
        gender, name, final_peak_yr = yr, peak_rank = rank 
    )

after_peak = national_w_rank |> 
    left_join(national_peak_yr, by=c("gender", "name")) |> 
    filter(yr >= final_peak_yr) |> 
    arrange(name, gender)


lowest_rank_after_peak = after_peak |> 
    arrange(name, gender, yr) |> 
    filter(rank == max(rank), .by=c(name, gender)) |> 
    filter(row_number() == max(row_number()), .by=c(name, gender)) |> 
    select(
        gender, name, bottom_yr = yr, bottom_rank = rank 
    )
    

name_change_rates = national_peak_yr |> 
    left_join(lowest_rank_after_peak, by=c("gender", "name")) |> 
    mutate(
        delta_years = bottom_yr - final_peak_yr,
        delta_rank = peak_rank - bottom_rank,
        change_rate = delta_rank/delta_years,
        )

name_change_rates |> 
    filter(peak_rank < 100) |> 
    filter(rank(change_rate)<10, .by="gender") |> 
    arrange(gender, change_rate)