How to reproduce Australian baby name counts yourself

If you need to check a claim about Australian baby name popularity, use the published ozbabynames dataset rather than a screenshot. According to the ozbabynames package documentation, the ozbabynames dataset contains popular Australian baby names by sex, state and year. Each row records one published count with the columns year, state, sex, name and count. That structure is enough to reproduce many checks: load the package, filter to the year, state, sex and name you care about, then read count.

The important limit is that this is published popular-name data, not a complete national births register. The documentation’s “Known Issues” section states that coverage is very uneven, that some states only provide very recent data, that some states only provide the top 50 or 100 names, and that the ACT does not provide counts, so no ACT data are included. South Australia has by far the best data, with full coverage of all names back to 1944. Any number you reproduce must be read with those limits attached.

What the dataset gives you

Fact: the package documentation says the bundled ozbabynames dataset has the columns year, state, sex, name and count.

Practical interpretation: each row is an aggregated published count, not a record of a single child. That means you can test claims such as “Charlotte was recorded 383 times for females in New South Wales in 2024”, but you cannot use the dataset to identify individuals or to recover names that a source did not publish.

Verify: check that your installed object has the same columns before you trust your own script.

Here are the first rows shown by head(ozbabynames) in the documentation:

Those rows are the package’s own example output. If your head(ozbabynames) call returns the same rows, your local copy is loaded correctly.

Install, load and inspect the data

The installation notes give two routes. The CRAN version can be installed with install.packages("ozbabynames"). The development version from GitHub is given as install_github("robjhyndman/ozbabynames").

A minimal reproducible start looks like this:

install.packages("ozbabynames")
# install_github("robjhyndman/ozbabynames")   # development version, per docs

library(ozbabynames)
head(ozbabynames)

If you plan to use the dplyr verbs shown in the documentation, load dplyr as well. The documented example also uses ggplot2 for charts and gganimate for one animated example, but you do not need those packages for counting.

The core counting recipe

For a single published count, filter to the exact row you need:

library(dplyr)

ozbabynames |>
  filter(year == 2024,
         state == "New South Wales",
         sex == "Female",
         name == "Charlotte")

This should return the row where count equals the published figure. If a headline gives a different number, your reproducible result is the filter above plus the returned count value.

For a “top names” claim, the documentation already supplies a reusable pattern. This version looks at 1952, groups by sex and name, sums counts across the states available in the data, orders from highest to lowest, and keeps the top 10 names within each sex:

ozbabynames_1952_top_10 <- ozbabynames |>
  filter(year == 1952) |>
  group_by(sex, name) |>
  summarise(count = sum(count)) |>
  arrange(-count) |>
  top_n(10) |>
  ungroup()

The practical rule is that you must state exactly what you filtered and summed. “Top 10” means nothing until you say the year, the sex, the included used one state or a sum of available states.

Aggregating across Australia: read this first

Because coverage is uneven, a cross-state sum is only the sum of rows present in ozbabynames. It is not automatically the same as every birth in Australia. The ACT is always absent. Some jurisdictions contribute only recent years, and some publish only a top 50 or 100 list, so rare names may be missing even when the count looks small. South Australia is the strongest historical source, with full coverage of all names back to 1944, but it is still one jurisdiction unless you deliberately restrict to it.

A safer method is to report either:

• the single state/sex/name/year count from the dataset; or
• an explicitly labelled sum, such as “sum of available ozbabynames counts for NSW, Victoria and Queensland in 2024, excluding the ACT”.

If you need a true jurisdictional total, verify against the original state source. The Sources section of the documentation lists South Australia, New South Wales, Tasmania, Victoria, Western Australia, Queensland and Northern Territory, includes three data.gov.au dataset links, and notes that only top 10 by rank are available after 2016.

Checklist before you trust a reproduced number

• Record the package version you used; the documentation page displays version 0.2.0.9000 and examples using 2024 data.
• Write the exact filter: year, state, sex, name.
• If you sum states, list every included jurisdiction and every known gap.
• Check whether the relevant state source is recent-only or top-50/100-only for that year.
• If you plot change over time, prefer one state and one sex so coverage changes do not look like name trends.

Known limits to record with your result

The documentation’s known issues are part of your result, not an afterthought:

• Coverage is very uneven across states and years.
• The ACT provides no counts, and no ACT data are included.
• Some states provide only very recent data.
• Some states provide only the top 50 or 100 names.
• South Australia has by far the best data, with full coverage of all names back to 1944.
• For the Tasmania source, the documentation notes that only top 10 by rank are available after 2016.

Next step: make the check citable

When you publish or share a reproduced count, include the script, the package version, the exact filter, the returned rows, and the date you accessed the data. That allows another researcher to rerun your steps and see whether the number still holds after later data updates.

If you later want a quicker, non-code way to explore names rather than rerun scripts, Australian Name has a matching tool at /match/. For reproducible research or verification, keep using the dataset method above and the source links on the package page.

General information only. This guide explains a public data method. It is not legal, tax, migration, credit, financial or personal advice, and it does not predict approvals, savings or outcomes. Australian Name is an independent publisher, not a births registry, government body, regulator, lender, broker or comparison service. The counts and coverage notes come from the ozbabynames documentation and its listed sources; check the current source before relying on any figure.