Gender prediction

A small utility based on genderize. Uses the Python client to the genderize web service (https://github.com/SteelPangolin/genderize).

Install

Create virtual environment:

$ python3 -m venv .venv

Activate virtual environment:

$ source .venv/bin/activate

Install required dependencies:

$ pip3 install -r requirements.txt

Run

Run the analysis:

$ python -m g names.txt

important The format of the input file should be:

unique_id   name

The unique_id is simply a number of identify each row that goes into the analysis

unique_id indiv_referees
1 Eva-Maria Mandelkow 
2 Jurgen Gotz
3 Karen Avraham
4 Stephen High
5 Ramanujan Hegde
6 Alfred Goldberg

Warning: There is a limit of 1000 requests per day!

Output

Output sample:

$ python -m g names.txt
name:Eva-Maria	gender:female	probability:1.0	count:6
name:Jurgen	gender:male	probability:0.97	count:36
name:Karen	gender:female	probability:1.0	count:5462
name:Stephen	gender:male	probability:1.0	count:2608
name:Ramanujan	gender:None
name:Alfred	gender:male	probability:1.0	count:230
name:Helene	gender:female	probability:1.0	count:255
name:Silvio	gender:male	probability:1.0	count:183
name:Junmin	gender:male	probability:1.0	count:2
name:Mark	gender:male	probability:1.0	count:6176
name:Henning	gender:male	probability:1.0	count:77
name:Ted	gender:male	probability:1.0	count:376
name:Olga	gender:female	probability:1.0	count:898
name:Mala	gender:female	probability:0.97	count:33
name:Thilo	gender:male	probability:1.0	count:11
results written to names_genders.txt

Usage directly from an eJP report file:

requires libraries dplyr, readr, mefa, argparse and XML

$ Rscript gender_parse.R -i names.txt -nlines 'number of lines that should be skipped when reading in the file'

arguments:
-i input file
-nlines number of lines to skip when reading in the eJP report file

Web application

Usage

cd to the directory 'webapp/'

$ python gender_webapp.py

Go to localhost:5001

Upload an eJP track report file. Results are automatically downloaded as a txt file.