Project 2: Top 5 Baby Names, 1926 to 2025

DATA 607 · Project 2 · Dataset 2 of 3

Author

Dillon Leeper

Published

October 11, 2026

Overview

This dataset comes from a classmate’s Discussion 5A post. It’s the Social Security Administration’s list of the five most popular baby names for girls and boys in every year from 1926 through 2025. The question from the post: which names stayed popular the longest, and do boys’ and girls’ names turn over at different rates?

Code

library(tidyverse)   # reading, reshaping, and charting
library(knitr)       # kable() for tables
library(scales)      # number formatting

# Shared look for all three Project 2 datasets
theme_p2 <- function() {
  theme_minimal(base_size = 12) +
    theme(
      plot.title       = element_text(face = "bold"),
      panel.grid.minor = element_blank(),
      legend.position  = "top"
    )
}
sex_colors <- c("Girls" = "#c9701a", "Boys" = "#2c5f8a")

Data source

Social Security Administration, Top 5 Names in Each of the Last 100 Years, accessed October 11, 2026. I copied the HTML table into ssa_top5_names_wide.csv without changing its layout. The page has a two-level header: a top row with “Girls” over the first five name columns and “Boys” over the last five, and a second row with “Year” and then “Rank 1” through “Rank 5” twice. I kept both header rows in the CSV so the code has to rebuild them.

The page notes that Michael was the top boys’ name 44 times and Mary the top girls’ name 30 times. My copy matches both counts, which is a quick check that nothing was lost in the copy.

Data structure before tidying

Code

csv_url <- paste0(
  "https://raw.githubusercontent.com/dillonleeper/",
  "DATA-607/main/assignments/week06/project2/baby-names/ssa_top5_names_wide.csv"
)

# Read every cell as text with no header, so both header rows come in as data.
# Try the GitHub copy first; fall back to the local copy if it can't be reached.
read_raw <- function(path) {
  read_csv(path, col_names = FALSE, col_types = cols(.default = col_character()))
}
raw_names <- tryCatch(read_raw(csv_url), error = function(e) read_raw("ssa_top5_names_wide.csv"))

head(raw_names, 5) |> kable(caption = "First five rows of the raw file, exactly as stored")

First five rows of the raw file, exactly as stored
X1	X2	X3	X4	X5	X6	X7	X8	X9	X10	X11
NA	Girls	NA	NA	NA	NA	Boys	NA	NA	NA	NA
Year	Rank 1	Rank 2	Rank 3	Rank 4	Rank 5	Rank 1	Rank 2	Rank 3	Rank 4	Rank 5
2025	Olivia	Charlotte	Emma	Amelia	Sophia	Liam	Noah	Oliver	Theodore	Henry
2024	Olivia	Emma	Amelia	Charlotte	Mia	Liam	Noah	Oliver	Theodore	James
2023	Olivia	Emma	Charlotte	Amelia	Sophia	Liam	Noah	Oliver	James	Elijah

The raw file has 102 rows and 11 columns: 2 header rows plus 100 years. Why it’s untidy:

Sex and rank are stored in the headers. They’re two variables, but they only exist as column labels.
The header is split across two rows. “Girls” and “Boys” appear once each, over blank cells, so R can’t name the columns on its own. Read normally, it would produce duplicate names like Rank 1 and Rank 1.
Each row holds ten observations. One row is one year, but it contains ten separate name-rank facts.

Transformation steps

Step 1: Rebuild the column names from the two header rows

The top header row only says “Girls” or “Boys” over the first column of each group, so fill() carries it across the blank cells. Then I combine it with the rank number into one snake_case name per column, like girls_rank_1.

Code

header <- tibble(
  sex  = unlist(raw_names[1, ]),
  rank = unlist(raw_names[2, ])
) |>
  fill(sex) |>   # "Girls" / "Boys" carried right over the blank cells
  mutate(col_name = if_else(
    is.na(sex),
    "year",
    paste(str_to_lower(sex), "rank", str_extract(rank, "\\d"), sep = "_")
  ))

header |> kable(caption = "How each column gets its new name")

How each column gets its new name
sex	rank	col_name
NA	Year	year
Girls	Rank 1	girls_rank_1
Girls	Rank 2	girls_rank_2
Girls	Rank 3	girls_rank_3
Girls	Rank 4	girls_rank_4
Girls	Rank 5	girls_rank_5
Boys	Rank 1	boys_rank_1
Boys	Rank 2	boys_rank_2
Boys	Rank 3	boys_rank_3
Boys	Rank 4	boys_rank_4
Boys	Rank 5	boys_rank_5

Step 2: Drop the header rows and apply the new names

Code

names_wide <- raw_names |>
  slice(-(1:2)) |>              # the two header rows are now in the names
  set_names(header$col_name)

head(names_wide, 3) |> kable()

year	girls_rank_1	girls_rank_2	girls_rank_3	girls_rank_4	girls_rank_5	boys_rank_1	boys_rank_2	boys_rank_3	boys_rank_4	boys_rank_5
2025	Olivia	Charlotte	Emma	Amelia	Sophia	Liam	Noah	Oliver	Theodore	Henry
2024	Olivia	Emma	Amelia	Charlotte	Mia	Liam	Noah	Oliver	Theodore	James
2023	Olivia	Emma	Charlotte	Amelia	Sophia	Liam	Noah	Oliver	James	Elijah

Step 3: Reshape wide to long

pivot_longer() turns the ten name columns into rows. Because each column name holds two variables, names_pattern pulls them apart: girls_rank_1 becomes sex = girls and rank = 1.

Code

names_tidy <- names_wide |>
  pivot_longer(
    cols      = -year,
    names_to      = c("sex", "rank"),
    names_pattern = "(girls|boys)_rank_(\\d)",
    values_to     = "name"
  ) |>
  mutate(
    year = as.integer(year),
    rank = as.integer(rank),
    sex  = str_to_title(sex),    # "girls" -> "Girls"
    name = str_squish(name)      # remove any stray spaces from the copy
  ) |>
  arrange(year, sex, rank)

head(names_tidy, 10) |> kable(caption = "Tidy data: one row per year, sex, and rank")

Tidy data: one row per year, sex, and rank
year	sex	rank	name
1926	Boys	1	Robert
1926	Boys	2	John
1926	Boys	3	James
1926	Boys	4	William
1926	Boys	5	Charles
1926	Girls	1	Mary
1926	Girls	2	Dorothy
1926	Girls	3	Betty
1926	Girls	4	Helen
1926	Girls	5	Margaret

Step 4: Check for missing or inconsistent values

There are no blanks in this table, but I check rather than assume. Every year should have exactly five girls’ names and five boys’ names, ranks 1 to 5, and no name should appear twice in the same year and sex.

Code

checks <- tibble(
  check  = c("Missing names",
             "Years without exactly 5 names per sex",
             "Same name twice in one year and sex",
             "Years covered"),
  result = c(
    sum(is.na(names_tidy$name) | names_tidy$name == ""),
    names_tidy |> count(year, sex) |> filter(n != 5) |> nrow(),
    names_tidy |> count(year, sex, name) |> filter(n > 1) |> nrow(),
    paste(min(names_tidy$year), "to", max(names_tidy$year))
  )
)
checks |> kable(col.names = c("Check", "Result"))

Check	Result
Missing names	0
Years without exactly 5 names per sex	0
Same name twice in one year and sex	0
Years covered	1926 to 2025

Code

stopifnot(
  !any(is.na(names_tidy$name)),
  all(count(names_tidy, year, sex)$n == 5)
)

All checks pass, so no rows needed fixing. The tidy table has 1000 rows (100 years × 2 sexes × 5 ranks) and 4 columns.

Analytical methods

All analysis uses names_tidy.

Streaks: for each name and sex, I sort the years it appeared in the top five and start a new streak whenever there’s a gap. A streak’s length is the number of consecutive years. A name can have more than one streak if it dropped out and came back.
Turnover: for each year after 1926, I count how many names in that year’s top five weren’t in the previous year’s top five. That’s the number of new entries. Summing new entries by sex, and by decade, shows how fast each list changes.
Variety: the number of different names that ever reached the top five for each sex.

No part of this analysis uses randomness, so no seed is needed.

Code

streaks <- names_tidy |>
  arrange(sex, name, year) |>
  group_by(sex, name) |>
  mutate(streak_id = cumsum(c(1, diff(year) != 1))) |>   # new id after any gap
  group_by(sex, name, streak_id) |>
  summarise(
    start     = min(year),
    end       = max(year),
    years     = n(),
    best_rank = min(rank),
    .groups   = "drop"
  )

turnover <- names_tidy |>
  group_by(sex, year) |>
  summarise(top5 = list(name), .groups = "drop") |>
  arrange(sex, year) |>
  group_by(sex) |>
  mutate(new_entries = map2_int(
    top5, lag(top5),
    \(now, before) if (is.null(before)) NA_integer_ else length(setdiff(now, before))
  )) |>
  ungroup()

Analysis 1: Which names stayed popular the longest?

Code

longest <- streaks |> arrange(desc(years), start) |> slice_head(n = 10)

longest |>
  transmute(
    Name             = name,
    Sex              = sex,
    `Streak (years)` = years,
    `Years`          = paste(start, "to", end),
    `Best rank`      = best_rank
  ) |>
  kable(caption = "The 10 longest consecutive runs in the top five")

The 10 longest consecutive runs in the top five
Name	Sex	Streak (years)	Years	Best rank
Michael	Boys	62	1949 to 2010	1
James	Boys	55	1926 to 1980	1
John	Boys	47	1926 to 1972	2
Robert	Boys	46	1926 to 1971	1
Mary	Girls	42	1926 to 1967	1
David	Boys	35	1950 to 1984	1
Christopher	Boys	27	1972 to 1998	2
Matthew	Boys	26	1981 to 2006	2
Joshua	Boys	26	1983 to 2008	3
William	Boys	24	1926 to 1949	4

Code

# Order names by the first year they appear, so the chart reads like a timeline
name_order <- names_tidy |>
  group_by(sex, name) |>
  summarise(first = min(year), .groups = "drop")

names_tidy |>
  left_join(name_order, by = c("sex", "name")) |>
  mutate(name = reorder(paste(name, sex), -first)) |>
  ggplot(aes(x = year, y = name, fill = rank)) +
  geom_tile(height = 0.8) +
  facet_wrap(~ sex, scales = "free_y") +
  scale_y_discrete(labels = \(x) str_remove(x, " (Girls|Boys)$")) +
  scale_fill_gradient(low = "#1f4463", high = "#bcd3e6", breaks = 1:5) +
  labs(
    title = "Every name that reached the top five, 1926 to 2025",
    x     = "Year",
    y     = NULL,
    fill  = "Rank"
  ) +
  theme_p2() +
  theme(axis.text.y = element_text(size = 8))

Code

top_boy  <- streaks |> filter(sex == "Boys")  |> slice_max(years, n = 1)
top_girl <- streaks |> filter(sex == "Girls") |> slice_max(years, n = 1)

Michael had the longest run of any name: 62 straight years in the boys’ top five (1949 to 2010). The longest girls’ run was Mary at 42 years (1926 to 1967). Nine of the ten longest streaks belong to boys’ names, and the four longest (Michael, James, John, and Robert) all lasted more than 45 years. The longest active streak is Emma’s: 24 years running, from 2002 through 2025.

Analysis 2: Do boys’ and girls’ names turn over at different rates?

Code

turnover_summary <- turnover |>
  group_by(sex) |>
  summarise(
    `Different names in top 5`  = n_distinct(unlist(top5)),
    `New entries, 1927 to 2025` = sum(new_entries, na.rm = TRUE),
    `Avg new entries per year`  = round(mean(new_entries, na.rm = TRUE), 2),
    `Years with no change`      = sum(new_entries == 0, na.rm = TRUE),
    `Median streak (years)`     = median(streaks$years[streaks$sex == first(sex)])
  ) |>
  rename(Sex = sex)

turnover_summary |> kable(caption = "Turnover in the top five, by sex")

Turnover in the top five, by sex
Sex	Different names in top 5	New entries, 1927 to 2025	Avg new entries per year	Years with no change	Median streak (years)
Boys	28	36	0.36	65	4
Girls	45	53	0.54	53	6

Code

turnover |>
  filter(!is.na(new_entries)) |>
  mutate(decade = paste0(year %/% 10 * 10, "s")) |>
  group_by(sex, decade) |>
  summarise(new_entries = sum(new_entries), .groups = "drop") |>
  ggplot(aes(x = decade, y = new_entries, fill = sex)) +
  geom_col(position = position_dodge(width = 0.8), width = 0.75) +
  scale_fill_manual(values = sex_colors) +
  scale_y_continuous(breaks = seq(0, 10, 2)) +
  labs(
    title    = "New names entering the top five, by decade",
    subtitle = "1920s covers 1927 to 1929 only; 2020s covers 2020 to 2025",
    x        = NULL,
    y        = "New entries (names)",
    fill     = NULL
  ) +
  theme_p2()

Code

girls_t <- turnover_summary |> filter(Sex == "Girls")
boys_t  <- turnover_summary |> filter(Sex == "Boys")
by_decade <- turnover |>
  filter(!is.na(new_entries)) |>
  mutate(decade = year %/% 10 * 10) |>
  group_by(sex, decade) |>
  summarise(n = sum(new_entries), .groups = "drop")
boys_2010s  <- by_decade |> filter(sex == "Boys",  decade == 2010) |> pull(n)
girls_2010s <- by_decade |> filter(sex == "Girls", decade == 2010) |> pull(n)

Girls’ names turn over faster. Over 100 years, 45 different girls’ names reached the top five, compared with 28 boys’ names. Girls’ lists had 53 new entries versus 36 for boys, and the boys’ top five stayed exactly the same from one year to the next in 65 of 99 years, against 53 for girls.

The decade chart shows that the gap isn’t constant. Through the 1970s, girls’ lists changed much more than boys’. Since the 1980s the pattern has mostly evened out, and in the 2010s it flipped: boys’ names had 9 new entries while the girls’ top five didn’t add a single new name (0 new entries). Emma, Olivia, Sophia, Isabella, and Ava held all five spots for the whole decade.

Conclusions

The most durable names are boys’ names. Michael stayed in the top five for 62 consecutive years, and four boys’ names lasted more than 45 years. The longest girls’ streak was Mary’s 42 years.
Girls’ names turn over more overall. 45 different girls’ names reached the top five, against 28 boys’ names, and girls had 53 new entries to boys’ 36.
The difference has narrowed. The gap came mostly from the mid-century decades. Recently it reversed, with a frozen girls’ top five in the 2010s while boys’ names churned.
Short-lived trends vs. lasting favorites: the median girls’ streak is 6 years and the median boys’ streak is 4. Boys have more short appearances (names that enter for a few years and leave), next to a handful of extremely long ones.

Limitations. The table only has ranks, not birth counts, so it can’t show how close the names were. A name at rank 5 might be just ahead of rank 6 or far ahead. Looking only at the top five also hides names that stayed popular just below it. A name like William can drop to rank 6 for a year and count as a broken streak.

AI Disclosure and Citation

Claude (Opus 5.5, accessed through claude.ai) helped with this dataset, including copying the SSA table into the CSV and checking it against the page’s own counts, writing the tidying and analysis code, drafting the write-up, and rendering the document to check the results.

Anthropic. (2026). Claude (Opus 5.5) [Large language model]. https://claude.ai/. Accessed October 11, 2026.