Names grouped by country

[djuxy]

Hi @philipperemy ,
Great work!

Is it possible to get a list of names (first/last) and % of appearance in a given country?

Cheers

[philipperemy]

@djuxy thank you :)

#14

Have a look at this one.

[djuxy]

I need a bit different output that i think i can not get from this one. :(

Something like this would work:

"France": { "Aimée": 0.000323, "Blanche": 0.000276, "Chanel": 0.000179, ... "Yves": 0.000123 }, "Mexico": { "Alejandro": 0.000422, "Francisco": 0.000376, "Juan": 0.000221, ... "Yolanda": 0.000087 },

[philipperemy]

@djuxy I'm generating you that!

[philipperemy]

@djuxy Here you go: https://drive.google.com/file/d/1wmVNXcfOYOcqhVesilI7LE0Lcm9-ivnw/view?usp=sharing.

It's a ZIP with this folder structure:

by_country
├── Afghanistan
│   ├── first_names.json
│   └── last_names.json
├── Albania
│   ├── first_names.json
│   └── last_names.json
├── Algeria
│   ├── first_names.json
│   └── last_names.json
├── Angola
│   ├── first_names.json
│   └── last_names.json
├── Argentina
│   ├── first_names.json
│   └── last_names.json

For a given JSON file, I provide pairs <name, count> where count is the number of occurrences of this name for this country. It's just a count without any normalization, except for a very simple filtering that I describe here:

NOTE: To save a bit of space and to reduce the noise, I kept the records with count>5

> cat France/first_names.json | head -n 20
{
  "A Marie": 8,
  "A-B": 6,
  "A-C": 17,
  "A-Charlotte": 7,
  "A-Claire": 9,
  "A-D": 8,
  "A-F": 7,
  "A-G": 6,
  "A-H": 8,
  "A-L": 28,
  "A-Laure": 19,
  "A-Lex": 7,
  "A-Line": 12,
  "A-Lise": 10,
  "A-Lys": 6,
  "A-Marie": 33,
  "A-S": 7,
  "A-So": 8,
  "A-Sophie": 30,

[djuxy]

@philipperemy Amazing! Thank You!

P.S. I don't need this, but probably somebody needs only male/female names. It would be great to separate male and female names if it's possible.

[philipperemy]

I'm happy I could help here.

Yeah def a great idea here. We have male/female in the FB dump so we could add a gender attribute ;)

I'll think of some nice way to integrate it

[philipperemy]

@djuxy I'm going to generate a giant CSV containing first_name,last_name,gender,country.

Then with pandas it will be easy for anyone to manipulate it and derive some stats or whatever metric they want.

[djuxy]

@philipperemy 👏 👏 👏 That would be great! Thank you!

[djuxy]

@djuxy I'm going to generate a giant CSV containing first_name,last_name,gender,country.

Then with pandas it will be easy for anyone to manipulate it and derive some stats or whatever metric they want.

I give a thought. It's better to have 2 csv:

• first one with: first_name, gender, country and count of appearance ,
• second one with: last_name, country and count of appearance

Why?
Gender is associated with first_name only.
We can easily see what is usage of first_name and last name separately in some country.

[philipperemy]

@djuxy I have just seen your message after I generated the large CSV. Have a look:

• https://drive.google.com/file/d/1wRQfw5EYpzulvRfHCGIUWB2am5JUYVGk/view?usp=sharing (~2GB tar bz2)

It is a folder containing one CSV per country using the country ISO code alpha_2 (hint: pycountry). I also added the country code in the last column if you wanted to concatenate all of them together. The uncompressed version takes around 10GB on the disk.

data
├── AE.csv
├── AF.csv
├── AL.csv
├── AO.csv
├── AR.csv
├── AT.csv

$ head FR.csv
Laure,Canet,F,FR
Louis,Givran,M,FR
Timothy,Dovin,M,FR
Anne Marie,Petiton,F,FR
Claudine,Solignac,F,FR
Florian,Burnat,,FR
Bendjy,Gobbo,M,FR
Danyel,Cambon,M,FR

It should contain all the information that you need.

The gender is either M, F or empty string (missing).

I agree that we can optimize it. Your suggestions make sense and I'll update all of that once I can find enough time.

Have a look in the mean time and let me know!

[SHEFKEVIN]

@philipperemy I just sent an email to your gmail :)

[philipperemy]

@SHEFKEVIN I replied you