I'd like to be able to determine if a person's name is likely to be a Japanese name, or likely to be a non-Japanese name.

Unless someone's compiled an exhaustive list of Japanese, and non-Japanese names, I suspect a good approach is to find a database that has a large list of Japanese people, and look at the frequency of names in it, and do the same with a database containing non-Japanese people.

The databases should contain female as well as male names, and should be in the English alphabet (romaji) or have kana spellings.

Is there any such data available?

I'd previously asked a similar question on Stack Overflow:
Determine whether a romanized name is Japanese or not, preferably in Ruby

  • As a rough Heuristic (that underfits), would just going though there names and seeing if they are made up entirely of Romanji letter pairs (/triples) for hiragana character, be sufficient? Dec 28 '13 at 0:18
  • @Oxinabox You would get many false positives like Kitae Dowon Misun Sunmi Jisu Miran Jae, Rosa Toni Anbessa Angela Anita Amanda... Aug 22 '14 at 15:19
  • 1
    Andrew, by "names" do you mean "given names"? Given names and family names would probably be 2 different databases justifying 2 different questions. Aug 22 '14 at 15:21
  • This "given name" list is extensive, and you can categorize each name by which country it is popular in. You can built two groups, Japanese and Non, and then count the frequency of each name in each group. opendata.stackexchange.com/a/5003/1511
    – philshem
    Aug 24 '17 at 10:09

Finally, Wikipedia and its sometimes tragical affection to compiling lists on everything and anything becomes handy:

  • But how and where do they get that data from?
    – Pacerier
    Jun 28 '15 at 8:22
  • Unfortunately these lists only contain 10 and 100 surnames, respectively, out of about 300k different Japanese surnames. Aug 19 '17 at 5:30

An exhaustive list of non-Japanese names would be an unwieldy thing! @ojdo's answer seems like a great start for the positive matches. For the United States, the best source for last names is probably the US Census Bureau, who, from their genealogy section, have a few kinds of name data:

  • 2010 and 2000: Excel file of 1000 most common surnames and ZIP with Excel and CSV of all surnames occurring more than 100 times. For both files, there are ranks, counts, and data about the racial distribution for the name.
  • 1990: Text files of proportion of surnames, male first names and female first names from a sample of survey responses. No race data, no absolute counts.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.