Google bias analysis with spaCy

spaCy

spaCy is a free, open-source library for advanced Natural Language Processing (NLP) in Python.

https://spacy.io/usage/spacy-101

Motivation

Google’s algorithms discriminate against women and people of colour
or plan to exterminate the White race?

We will try to check why Google's image search results for the query 'happy white woman' differ from other search engines. We will use results from Bing for comparison. Results are limited by dates from 2010-01-01 to 2019-01-01.

Google Explains Why Some Image Search Results Look Racist

Dataset

data/dataset.csv

title - image title
link - url to image
baseline - search result position
query - search query
engine - search engine google/bing

Notebooks

1. Similarity analysis with spaCy
2. Dependency parsing with spaCy
3. Transformers with spaCy