Methodology: where the numbers and notes come from

This page explains where every number and every explanation on the site comes from. If you want to know what sits behind a figure or an origin note, the answer is here.

Where do the numbers come from?

Surname counts come from TÜİK’s Nüfus İsim Portalı, the name portal of the Turkish Statistical Institute. It publishes the most common surnames from the Address-Based Population Registration System (ADNKS). We hold two lists for every year from 2018 to 2025: the national top 30 and the top 30 for each of the 81 provinces. Every row has a count.

To turn counts into rates we use the ADNKS population tables from the same portal. They give the national total and the population of each province, year by year.

The second register is NVİ, the General Directorate of Population and Citizenship Affairs at the Ministry of the Interior. NVİ surname figures only appear when they are given to the press. Our one NVİ source is an Anadolu Ajansı report of 17 July 2021.

We read Anadolu Ajansı’s ADNKS reports (11 March 2025 and 11 May 2026) as a cross-check only. Their top-three counts match the portal exactly. They are not entered as separate rows.

Why do ADNKS and NVİ give different numbers?

They count different populations.

ADNKS is TÜİK’s address-based register. It covers people resident in Türkiye on 31 December of each year. Foreign nationals with residence or work permits are included. According to TÜİK’s notes, foreigners on visas or permits shorter than 90 days are not, and nor are Syrians under temporary protection. People living in dormitories, barracks, prisons or care homes are counted where the institution is.

NVİ figures come from the civil register: the family registers (aile kütükleri) kept under the Population Services Law, No. 5490. A record there does not close when a person dies. The death is entered on the record as an event. Citizens living abroad also report births, marriages and deaths through consulates. The 2021 report does not say which records were counted, so we cannot say for certain why the totals differ.

The gap is large. For 2021, NVİ gives 1,580,313 people named Yılmaz; ADNKS gives 1,160,248. Kaya: 1,154,158 against 875,113. Demir: 1,105,381 against 849,475. That pattern fits two registers with different coverage. So we never put the two in one chart, and every figure carries a label such as “TÜİK ADNKS 2025” or “NVİ”.

After its top three, the NVİ report names seven more surnames: Çelik, Şahin, Yıldız, Yıldırım, Öztürk, Aydın and Özdemir. It gives no counts and does not say the list is in order. We do not turn it into ranks.

Which date do the figures refer to?

For ADNKS, the reference date is 31 December of the year. “ADNKS 2025” means the people resident in Türkiye on 31 December 2025.

The NVİ report was published on 17 July 2021 and gives no data date. We store and show those figures as 2021.

How are per-1,000, “1 in N” and the classroom maths worked out?

Every calculation uses the ADNKS surname count and the ADNKS population for the same year. Example: in 2025, 1,177,896 people were named Yılmaz, out of a population of 86,092,168.

• Per 1,000 = count ÷ population × 1,000. For Yılmaz: 1,177,896 ÷ 86,092,168 × 1,000 ≈ 13.68.
• 1 in N = population ÷ count, rounded to the nearest whole number. For Yılmaz: 86,092,168 ÷ 1,177,896 ≈ 73, so one person in 73.
• Classroom maths = n × count ÷ population, the expected number of bearers in a group of n people. In a group of 100 you would expect about 1.4 people named Yılmaz; in a group of 1,000, about 14.

The classroom figure is an average. A real school or class will differ by province and neighbourhood.

How are share of the top 10 and rank change calculated?

• Share of the top 10 = count ÷ the combined count of that year’s top 10 surnames × 100. In 2025 the top 10 add up to 7,448,820. Yılmaz’s share is 1,177,896 ÷ 7,448,820 × 100 ≈ 15.8%.
• Rank change = last year’s rank − this year’s rank. A positive number means the surname climbed; a negative one means it fell. If a surname was not on last year’s list, no rank change is shown.
• Change since the first year = latest count − count in the first year of the series.

Surnames are inherited, not chosen. That is why ranks barely move from one year to the next. Counts change only through births, deaths, migration and the occasional name change.

What does “rank without count” mean?

Some sources give a surname’s rank but not how many people carry it. In that case the page shows the rank only. We never estimate the missing count or fill it in from another source. Our data model allows the count to be empty.

All of TÜİK’s top-30 lists on the portal come with counts. A rank without a count turns up in press reports.

What happens when two surnames have the same count?

We keep TÜİK’s ranks exactly as published. Surnames with the same count share a rank, and the next rank is not skipped. A “top 30” list can therefore have more than 30 rows. Some province lists run to 33.

What do the province figures cover?

For each province we use TÜİK’s 30 most common surnames, with counts, for every year from 2018 to 2025. If a surname is missing from a province’s list, it is not absent from that province. It is simply below that province’s top 30.

Neighbouring provinces are derived from Natural Earth boundary data. Region names were checked against Wikipedia’s list of provinces.

How are full-name figures estimated?

There is no official answer to “how many Mehmet Yılmaz are there?”. Full-name figures on the site are therefore estimates:

Estimate = P(given name) × P(surname) × population

P(given name) is the given name’s share of the population, taken from the TÜİK given-name data used by our sister site, turkiyenames.com. P(surname) is the surname count divided by the population. The formula assumes that given names and surnames are independent of each other. They are not: naming varies by region, generation and family custom. Each full-name page says plainly that its figure is an independence estimate.

How are meanings and origins researched?

Meanings rest on dictionary sources:

• TDK’s Güncel Türkçe Sözlük confirms the meaning when a surname is also an ordinary word (Demir, Kaya, Yılmaz).
• Root-language tags follow the conventions of Nişanyan Sözlük: Turkic, Arabic, Persian, Greek, Armenian, Kurdish, Mongolic, French.
• The 1934 Surname Law and Surname Regulation show the rules surnames were formed under.
• Wikidata supplies surname records and notable people who carry them.

Wiktionary glosses are used as hints only. We copy no text from any source. Every meaning is written in our own words.

How do you decide between established, probable and disputed?

Every origin note carries one of three levels:

• Established: a dictionary (TDK or an etymological dictionary in the Nişanyan tradition) gives the meaning, and nobody seriously disputes it. Examples: Kaya, Demir, Yılmaz, Demirci, Konyalı.
• Probable: the analysis is transparent, but no source deals with this exact surname, or one element has two standard readings. Most compounds (Akdoğan) and most names ending in -kan sit here.
• Disputed: two or more sources give different etymologies, or the popular explanation clashes with the dictionary one. We then give each view in a sentence, with its source.

Folk etymology is never presented as fact. If we mention one, we say “a popular explanation is …” and put the documented one beside it.

Some surnames appear under “1934 coinages”. That is a structural label. It covers forms that are not dictionary headwords, are not used as given names, and are built from the stock elements of 1934 (Öz-, Ak-, Gök-, -soy, -kan and the like). It does not claim that a particular surname was coined in 1934.

How do you handle sensitive topics?

The linguistic origin of a word tells you nothing about the ancestry of the people who carry the name. That is the site’s first rule.

• We never attribute ethnicity, religion, sect or political views to the bearers of a surname.
• We do not write sentences of the form “this surname is common among [group]”.
• A Kurdish, Armenian, Greek, Arabic or Persian root is stated only as a dictionary fact about the word, with a source.
• Tribal names (Kayı, Avşar, Çepni) are described as historical names. We never say bearers descend from the tribe.

How often is the data updated?

Once a year. TÜİK usually releases the ADNKS results in February. The surname figures tend to follow in February or March, sometimes later: the 2024 figures reached the press on 11 March 2025, the 2025 figures on 11 May 2026. When a new year is published, every calculation is rebuilt.

Meaning and origin notes grow continuously. Each page shows the date it was last updated.

What happens if you find a mistake?

We check every report against its source. Confirmed errors are fixed, and material corrections are listed with date, page and change on the Corrections page. The full list of sources is on the Sources page.