bubsname
Search

transparency · open sources · 95 years of data

Where do these numbers come from?

Every figure on bubsname.com is sourced from publicly-licensed government registries — primarily the eight Australian state and territory Registries of Births, Deaths & Marriages, plus Wikipedia and Wikidata for etymology and multilingual variants. All licences allow free re-use; this page documents what we have, how we aggregate it, and what we don't have.

State & territory baby name registries

Australia has no federal baby names dataset — there is no equivalent of the US Social Security Administration list. Each of the eight jurisdictions publishes its own annual rankings, in different formats, with different depth and on different schedules. We aggregate them into a single national view.

Primary source for national aggregation: NSW Registry of Births, Deaths & Marriages — Popular Baby Names — the deepest and longest-running of our datasets (1952–present). Western Australia's hidden JSON API extends our backbone further back, to 1930, though only for the top 10 names per year.

National aggregation method

For each (year, gender) combination we compute the national ranking as follows:

  1. For each state, we have (name, year, gender, count, state_rank).
  2. We sum count across all states that published data for that (name, year, gender) tuple.
  3. We re-rank by the summed count, breaking ties alphabetically.

States that publish only rankings (no counts) — currently TAS for 2017+ and ACT for all years — do not contribute to the national sum, but their per-state rankings still show on the per-state pages. The national figure is therefore a lower bound on actual registrations: real births from TAS and ACT exist, we just don't have magnitudes.

For years before a particular state began publishing, that state is simply absent from the sum. Older years (1930s–1950s) are dominated by NSW and WA because other states started later — see the per-state ranges in §1.

Enrichment: Wikipedia & Wikidata

Etymology, origin language family, notable bearers and multilingual variants come from two open knowledge sources, processed for our top 200 names by overall registrations:

  • Wikipedia (English): https://en.wikipedia.org/ — article intro paragraphs and origin labels (Greek, Latin, Hebrew, Old English, Irish, Germanic, etc.). Licence: CC BY-SA 4.0. We attribute on each name page where a Wikipedia excerpt is used.
  • Wikidata (SPARQL endpoint): https://www.wikidata.org/ — multilingual variants across 21+ languages (Vietnamese, Mandarin, Arabic, Greek, Italian and so on — selected for relevance to Australia's largest ancestry groups per ABS Census 2021). Licence: CC0 1.0 (public domain, no attribution required).

Coverage of enrichment data is approximately 95% of our top 200 (≈190 names from Wikipedia and ≈175 from Wikidata). Long-tail names without Wikipedia articles show registry statistics only, with no etymology section.

NZ comparator

For the Australia vs New Zealand page we use the NZ Department of Internal Affairs name dataset (1954–2023), mirrored as a CSV by the maintainers of the nzbabynames R package on GitHub. The DIA data is aggregated (Sex=NA), so cross-Tasman gender split is inferred from the AU side via case-insensitive name matching.

Licence: CC BY 4.0 NZ. The comparator counts a name as "in both top 30" only when it is actually ranked top-30 in both countries in the chosen year — not just present somewhere in either dataset.

Refresh cadence

State BDM datasets typically refresh in January–March each year for the previous year's data. We re-ingest within a week of publication. There is no automatic crawl — we run the ingest scripts manually so each refresh can be sanity-checked.

Wikipedia and Wikidata enrichment is run quarterly or when new top-200 names appear in the registry data. Both ingest scripts include a guard that skips re-fetching names that already have content.

How current is the data right now? Each state page lists the years actually present in our database, and per-name pages show the most recent year on record. If you need specific currency confirmation for a citation or article, drop us a line.

Known coverage gaps

  • Northern Territory — the NT Registry of Births, Deaths & Marriages has not published baby name rankings since 2022. Our most recent NT data is therefore from 2022; we will refresh if and when publication resumes. Wayback Machine snapshots of the NT government page confirm the publication hiatus.
  • Tasmania (2017 onwards) — TAS publishes its top 10 rankings only, without counts. Per-state pages for 2017+ TAS show ranks but no magnitude bars, and TAS does not contribute to the national sum for those years. Historical 2010–2016 data has full counts via the Tasmanian Government open data archive (mirrored on GitHub by Prof. Rob Hyndman, CC BY 4.0).
  • Australian Capital Territory — ACT does not publish a structured dataset, only annual December press releases. We have top 10 ranks for 2020, 2021, 2023, 2024 and 2025. Some years (2018, 2019, 2022) are missing because those press releases listed only the top 5 in prose rather than a full top 10.
  • Western Australia — WA BDM publishes only the top 10 per year via an undocumented JSON endpoint (1930 onwards). No long-tail counts beyond the top 10 are available. Long-term WA trends therefore reflect only the most popular names, not the full distribution.
  • South Australia (pre-2014) — the SA CKAN dataset starts in 2014. A legacy 1944–2013 archive exists on data.sa.gov.au in ZIP form; we have not yet ingested it. This is a known follow-up.
  • Pre-1952 (other than WA) — for years before 1952, only WA is available. Aggregate national rankings for the 1930s and 1940s therefore reflect WA only.

Editorial standards

We follow a few simple rules to keep editorial overlays honest:

  • No value judgments about individuals. "Extinct names" and "rising names" describe registry patterns, not opinions about people who carry those names today. We avoid characterising any name as "good", "bad", "outdated" or similar.
  • Honest framing of sparse data. Names that fell below their jurisdiction's top-100 threshold show as "last ranked in YYYY" rather than "extinct since YYYY" — the registry simply stops publishing them, even if babies are still being given those names privately.
  • Plural and tense awareness. "1 baby" (singular) where there is one; "last ranked" (past tense) for names that have fallen out of top-100; "today" only when the most recent data point is within the current or previous year.
  • Predictions are explicitly modelled. The predictions page uses a transparent linear regression on 2020–2024 data, not opaque ML. Confidence is bounded by the assumption that recent trends continue.
  • Citations on every claim. Per-name pages cite Wikipedia article URLs; per-state pages link to the originating BDM dataset; the national figure is reproducible from the per-state numbers on this site.

Indigenous names — protocols

bubsname acknowledges the Traditional Custodians of the lands across Australia and pays respect to Elders past and present. Indigenous Australian naming connects to mob, land and identity in ways that the colonial BDM registry format cannot adequately represent.

We follow the AIATSIS Code of Ethics for Aboriginal and Torres Strait Islander Research as guidance. Specifically:

  • We do not redistribute curated Indigenous-specific name lists. For community-sourced material, please refer to AIATSIS family-history resources and to land-council and language-centre publications.
  • Names recorded in 19th- and early-20th-century BDM data may not reflect community preference and were sometimes imposed during the assimilation era. We surface this context where we display historical rankings.
  • We respect "sorry business" protocols. If a name on the Service should be removed, withheld from search or contextualised differently following the death of a community member, contact us — we respond within 7 days.

Warning: Aboriginal and Torres Strait Islander people are advised that this site may include the names of people who have died. We have not been able to filter for "sorry business" individually; we rely on community feedback to do so.

Corrections

Every number on bubsname.com is verifiable against its original source — that is the point of linking every dataset back to the publishing registry. If something doesn't match what you know, you are likely right and we are likely behind on an ingest.

Get in touch — describe what you saw, where you saw it, and the source you're comparing against. We aim to triage within 48 hours and either correct the issue or explain why what we show reflects the source data.

For privacy enquiries or data-subject requests, see our Privacy policy.