AE 13: Debugging R code

Packages

We will use the following packages in this application exercise.

• tidyverse: For data import, wrangling, and visualization.
• babynames: For working with the Social Security Administration’s baby names data.

library(tidyverse)
library(babynames)

Popular baby names

The Social Security Administration maintains detailed historic records on every child born in the United States dating back to 1880.¹ The data is published annually at the national and state-levels, and includes every name excluding those with fewer than 5 occurrences. The babynames package provides a convenient interface to this data. ²

¹ Source: SSN

² The package is no longer actively maintained. We will use a forked version of the package which has been submitted as a pending pull request.

The instructor has written code to analyze the popularity of baby names in the United States. Alas, the code is not working correctly. Your task is to debug the code and fix any issues. Pay attention to any errors, warnings, or messages generated by the code. It is possible for the code to run without any notification from R, but still produce incorrect output.

Write a function to show trends over time for specific name

Your turn: Copy the reprex from the course discussion board. Turn this into clean code using reprex::reprex_clean() and add it to the code chunk below. Then debug the code, fix any issues, and post the corrected code on the discussion board using a reprex.

# add code here

Compare naming trends to Disney princesses around film release years

Your turn: Fix the code below to show the popularity over time of baby names that are shared with Disney princesses.

# create data frame of disney princess films
disney <- tribble(
  "princess",  "film", "release_year",
  "Snow White", "Snow White and the Seven Dwarfs", 1937,
  "Cinderella", "Cinderella", 1950,
  "Aurora", "Sleeping Beauty", 1959,
  "Ariel", "The Little Mermaid", 1989,
  "Belle", "Beauty and the Beast", 1991,
  "Jasmine", "Aladdin", 1992,
  "Pocahontas", "Pocahontas", 1995,
  "Mulan", "Mulan", 1998,
  "Tiana", "The Princess and the Frog", 2009,
  "Rapunzel", "Tangled", 2010,
  "Merida", "Brave", 2012,
  "Elsa", "Frozen", 2013,
  "Moana", "Moana", 2016
)

Error in `tribble()`:
! Must specify at least one column using the `~name` syntax.

# join together the data frames
babynames |>
  # ignore men named after princesses - is this fair?
  filter(sex == F) |>
  inner_join(disney, by = c("name" = "princess")) |>
  mutate(name = fct_reorder(.f = name, .x = release_year)) |>
  # plot the trends over time, indicating release year
  ggplot(mapping = aes(x = year, y = n)) +
  facet_wrap(facets = vars(name, film), scales = "free_y", labeller = label_both()) +
  geom_line() +
  geom_vline(mapping = aes(xintercept = release_year), linetype = 2, alpha = .5) +
  scale_x_continuous(breaks = c(1880, 1940, 2000)) +
  theme_minimal() +
  labs(title = "Popularity of Disney princess names",
       x = "Year",
       y = "Number of births")

Error in eval(expr, envir, enclos): object 'disney' not found

Write a function to show trends over time for the top N names in a specific year

Your turn: Modify the function so that if n_rank is greater than 6, it is capped at 6 and a message is displayed alerting the user to this change.

top_n_trend <- function(n_year, n_rank = 5) {
  # create lookup table
  top_names <- babynames |>
    summarize(count = as.numeric(sum(n)), .by = c(name, sex)) |>
    filter(count > 1000) |>
    select(name, sex)

  # filter babynames for top_names
  filtered_names <- babynames |>
    inner_join(top_names, by = join_by(sex, name))

  # get the top N names from n_year
  top_names <- filtered_names |>
    filter(year == n_year) |>
    summarize(count = sum(n), .by = c(name, sex)) |>
    group_by(sex) |>
    mutate(rank = min_rank(desc(count))) |>
    filter(rank <= n_rank) |>
    arrange(sex, rank) |>
    select(name, sex, rank)

  # keep just the top N names over time and plot
  filtered_names |>
    inner_join(select(top_names, sex, name), by = join_by(sex, name)) |>
    ggplot(mapping = aes(x = year, y = n, color = name)) +
    facet_wrap(facets = vars(sex), ncol = 1) +
    geom_line() +
    scale_color_brewer(type = "qual", palette = "Set3") +
    labs(
      title = str_glue("Most Popular Names of {n_year}"),
      x = "Year",
      y = "Number of births",
      color = "Name"
    ) +
    theme_minimal()
}

top_n_trend(n_year = 1986)

top_n_trend(n_year = 2014)

top_n_trend(n_year = 1986, n_rank = 10)

Warning in RColorBrewer::brewer.pal(n, pal): n too large, allowed maximum for palette Set3 is 12
Returning the palette you asked for with that many colors

Warning: Removed 1060 rows containing missing values or values outside the scale range
(`geom_line()`).