You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think this is pretty universal: some names and surnames are more common than others. It would be a bit more realistic if we could assign weights to names and surnames and imitate their real distribution.
Sounds like a good feature, but it may be a bit complicated to implement in terms of finding a source that can cover the surnames and their associated weights of all the nationalities.
I've found official stats about names and surnames in Spain, and I've assembled weighted lists for male and female names, as well as surnames:
Antonio 38 Maria Carmen 38 Garcia 57
Manuel 34 Maria 34 Rodriguez 36
Jose 33 Carmen 21 Gonzalez 36
Francisco 28 Ana Maria 16 Fernandez 36
David 22 Maria Pilar 15 Lopez 34
Juan 20 Laura 15 Martinez 33
Javier 19 Josefa 15 Sanchez 32
Jose Antonio 18 Isabel 15 Perez 31
Daniel 18 Maria Dolores 15 Gomez 19
Francisco Javier 17 Maria Teresa 14 Martin 19
... ... ...
Emilio Jose 1 Elizabeth 1 Asensio 1
Jose Andres 1 Meritxell 1 Reina 1
Simon 1 Desiree 1 Polo 1
Luis Antonio 1 Gregoria 1 Ojeda 1
1000 TOTAL Antonia Maria 1 Ramon 1
... ...
Maria Manuela 1 Carrera 1
Mia 1 Toledo 1
Maria Candelaria 1 Ayala 1
Maria Gracia 1 Alcaraz 1
1000 TOTAL Hernando 1
...
Mejias 1
Carvajal 1
Rosales 1
Toro 1
1000 TOTAL
After looking at the API code, it looks like simply using a list with 38 "Antonio" lines, 34 "Manuel", etc. would work without any code change.
Activity
keitharm commentedon May 16, 2022
Sounds like a good feature, but it may be a bit complicated to implement in terms of finding a source that can cover the surnames and their associated weights of all the nationalities.
atnbueno commentedon May 17, 2022
What I had in mind was a backwards compatible solution:
Solution #1: allow numbers after the names/surnames, and if present, use them as weight (no number meaning weight=1)
Solution #2: allow repeated names in the lists (equivalent to integer weights) and just pick one random line
If there's no weight information, the data stays as is. If someone finds a weighted source for a particular version, they can do a PR.
atnbueno commentedon Dec 1, 2022
I've found official stats about names and surnames in Spain, and I've assembled weighted lists for male and female names, as well as surnames:
After looking at the API code, it looks like simply using a list with 38 "Antonio" lines, 34 "Manuel", etc. would work without any code change.
Should I do a PR with such lists?