A very simple API to get gender probability on a name and/or article.

It uses Stanford NER, version 3.2.0. Stanford NER is a package that provides a high-performance machine learning based named entity recognition system, incluidng facilities to train models from supervised training data and pre-trained models for English. The software is licensed under full GPL. You can find the software in lib/ner/

The name's gender is recognized based on estatical census data from US and UK.

USAGE

This is a sinatra app. To run it locally:

'bundle install'
'ruby app.rb'

API

It returns json.

Guess gender of the name.

GET gender?name=anna


{
    name: "anna",
    gender: "female"
}

Get gender of a text.

POST content?text=Something that he wanted to say.

{
    "result": "Male",
    "counts": {
        "male": 7,
        "female": 0,
        "neutral": 0
    }
}

Get how many femail and male are in a text

POST people?text=Maria and Anne were talking.


{
    "Maria": {
        "first": null,
        "surname": "Maria",
        "count": 1,
        "gender": "female"
    },
    "Anne": {
        "first": null,
        "surname": "Anne",
        "count": 1,
        "gender": "female"
    }
}

LICENSE

The software is licensed under the GPL. Please see the file LICENCE.txt

CONCEPTION

This API was concieved during MozFest 2013 London, at Nathan Matia's 3 hours workshop "Measuring the News: Tracking Content and Engagement". Notes for the session on https://festival.etherpad.mozilla.org/journalism_measuring-the-news-tracking-content-and-engagement