name2nat: a Python package for nationality prediction from a name

name2nat is a Python package that predicts the nationality of any name written in Roman letters. For example, it returns the correct output Korean for my name `Kyubyong Park'. Needless to say, it is not possible to guess somebody's nationality 100% right from their name. After all, nationality can change, you know. However, it is also true that there is a tendency between names and nationality. So it turns out statistical classifiers for this task works to some extent. Details are explained below.

NaNa Dataset

Construction

I constructed a new dataset for this project because I failed to find any available dataset that is big and comprehensive enough.

• STEP 1. Downloaded and extracted the 20200601 English wiki dump (enwiki-20200601-pages-articles.xml).
• STEP 2. Iterated all pages and collected the title and the nationality. I regarded the title as a person if the Category section at the bottom of each page included ... births (green rectangule), and identified their nationality from the most frequent nationality word in the section (red rectangules).

* STEP 3. Randomly split the data into train/dev/test in the ratio of 8:1:1 within each nationality group.

Stats

Nationality	Train	Dev	Test
Total 1112902	890248	111286	111368
Afghan	778	97	98
Albanian	2193	274	275
Algerian	1592	199	200
American	241772	30221	30222
Andorran	188	24	24
Angolan	504	63	63
Argentine	8926	1116	1116
Armenian	1600	200	201
Aruban	93	12	12
Australian	40536	5067	5067
Austrian	9192	1149	1149
Azerbaijani	1331	166	167
Bahamian	233	29	30
Bahraini	237	30	30
Bangladeshi	1636	204	205
Barbadian	372	47	47
Basque	961	120	121
Belarusian	2338	292	293
Belgian	7907	988	989
Belizean	148	19	19
Beninese	199	25	25
Bermudian	270	34	34
Bhutanese	144	18	18
Bolivian	657	82	83
Bosniak	81	10	11
Botswana	252	31	32
Brazilian	11234	1404	1405
Breton	118	15	15
British	45922	5740	5741
Bruneian	115	14	15
Bulgarian	3926	491	491
Burkinabé	289	36	37
Burmese	944	118	118
Burundian	140	17	18
Cambodian	360	45	46
Cameroonian	1028	129	129
Canadian	34152	4269	4270
Catalan	1717	215	215
Chadian	139	17	18
Chilean	2838	355	355
Chinese	9494	1187	1187
Colombian	2620	328	328
Comorian	54	7	7
Congolese	35	4	5
Cuban	1938	242	243
Cypriot	1016	127	128
Czech	7244	906	906
Dane	32	4	5
Djiboutian	54	7	7
Dominican	1580	198	198
Dutch	14916	1864	1865
Ecuadorian	874	109	110
Egyptian	2776	347	348
Emirati	621	78	78
English	77159	9645	9645
Equatoguinean	193	24	25
Eritrean	133	17	17
Estonian	2028	254	254
Ethiopian	733	92	92
Faroese	284	35	36
Filipino	3928	491	491
Finn	68	8	9
French	40841	5105	5106
Gabonese	180	23	23
Gambian	220	28	28
Georgian	262	33	33
German	42388	5299	5299
Ghanaian	2036	255	255
Gibraltarian	98	12	13
Greek	5975	747	747
Grenadian	139	17	18
Guatemalan	563	70	71
Guinean	584	73	74
Guyanese	358	45	45
Haitian	561	70	71
Honduran	500	63	63
Hungarian	7220	903	903
I-Kiribati	40	5	6
Indian	22692	2836	2837
Indonesian	2820	352	353
Iranian	5010	626	627
Iraqi	1252	157	157
Irish	11844	1481	1481
Israeli	5149	644	644
Italian	29336	3667	3668
Jamaican	1422	178	178
Japanese	21216	2652	2652
Jordanian	490	61	62
Kazakh	24	3	4
Kenyan	1609	201	202
Korean	7896	987	988
Kuwaiti	396	50	50
Kyrgyz	16	2	2
Lao	26	3	4
Latvian	1693	212	212
Lebanese	1246	156	156
Liberian	294	37	37
Libyan	271	34	34
Lithuanian	1979	247	248
Macedonian	1099	137	138
Malagasy	232	29	29
Malawian	219	27	28
Malaysian	2582	323	323
Maldivian	152	19	20
Malian	385	48	49
Maltese	663	83	83
Manx	150	19	19
Marshallese	32	4	4
Mauritanian	96	12	12
Mauritian	263	33	33
Mexican	8648	1081	1081
Moldovan	1000	125	125
Mongolian	504	63	64
Montenegrin	955	119	120
Moroccan	1457	182	183
Mozambican	210	26	27
Namibian	588	74	74
Nauruan	32	4	4
Nepalese	773	97	97
Nicaraguan	285	36	36
Nigerian	4060	507	508
Nigerien	143	18	18
Norwegian	13512	1689	1690
Omani	197	25	25
Pakistani	3762	470	471
Palauan	35	4	5
Palestinian	528	66	66
Panamanian	474	59	60
Paraguayan	1012	127	127
Peruvian	1521	190	191
Portuguese	4734	592	592
Qatari	548	68	69
Romanian	6551	819	819
Russian	21274	2659	2660
Rwandan	269	34	34
Salvadoran	507	63	64
Sammarinese	198	25	25
Samoan	596	75	75
Saudi	1496	187	188
Senegalese	823	103	103
Serb	44	6	6
Singaporean	1316	165	165
Slovak	2867	358	359
Slovene	88	11	12
Somali	116	14	15
Sotho	49	6	7
Sudanese	348	44	44
Surinamese	200	25	25
Swazi	114	14	15
Syriac	78	10	10
Syrian	1047	131	131
Taiwanese	1946	243	244
Tajik	61	8	8
Tamil	1399	175	175
Tanzanian	627	78	79
Thai	2747	343	344
Tibetan	265	33	34
Togolese	211	26	27
Tongan	456	57	57
Tunisian	1072	134	134
Turk	79	10	10
Tuvaluan	66	8	9
Ugandan	1052	132	132
Ukrainian	6198	775	775
Uruguayan	2267	283	284
Uzbek	62	8	8
Vanuatuan	116	15	15
Venezuelan	1937	242	243
Vietnamese	1257	157	158
Vincentian	8	1	1
Welsh	5270	659	659
Yemeni	322	40	41
Zambian	510	64	64

Downloadable Link

• You can download the dataset here.

name2nat

Installation

pip install name2nat

Usage

>>> from name2nat import Name2nat

>>> my_nanat = Name2nat()

>>> names = ["Donald Trump", # American
         "Moon Jae-in", # Korean
         "Shinzo Abe", # Japanese
         "Xi Jinping", # Chinese
         "Joko Widodo", # Indonesian
         "Angela Merkel", # German
         "Emmanuel Macron", # French
         "Kyubyong Park", # Korean
         "Yamamoto Yu", # Japanese
         "Jing Xu"] # Chinese
>>> result = my_nanat(names, top_n=3)
>>> print(result)
# (name, [(nationality, prob), ...])
# Note that prob of 1.0 indicates the name exists
# in Wikipedia.
[
('Donald Trump', [('American', 1.0)])
('Moon Jae-in', [('Korean', 1.0)])
('Shinzo Abe', [('Japanese', 1.0)])
('Xi Jinping', [('Chinese', 1.0)])
('Joko Widodo', [('Indonesian', 1.0)])
('Angela Merkel', [('German', 1.0)])
('Emmanuel Macron', [('French', 1.0)])
('Kyubyong Park', [('Korean', 0.9985014200210571), ('American', 0.000289416522718966), ('Bhutanese', 0.00025851925602182746)])
('Yamamoto Yu', [('Japanese', 0.7050493359565735), ('Taiwanese', 0.12779785692691803), ('Chinese', 0.04263153299689293)])
('Jing Xu', [('Chinese', 0.8626819252967834), ('Taiwanese', 0.09901007264852524), ('American', 0.022995812818408012)])
]

Training

I use a powerful NLP library Flair to train a text classifier model. A bidirectional GRU layer is employed.

python train.py

Evaluation

python predict.py;
python eval.py --gt nana/test.tgt --pred test.pred

Results

K	Precision@K
1	61310/111368=55.1
2	77480/111368=69.6
3	86703/111368=77.9
4	92491/111368=83.0
5	96697/111368=86.8

References

If you use this code for research, please cite:

@misc{park2018name2nat,
  author = {Park, Kyubyong},
  title = {name2nat: a Python package for nationality prediction from a name},
  year = {2020},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/Kyubyong/name2nat}}
}