Code for Twitter Algorithmic Bias Challenge (https://hackerone.com/twitter-algorithmic-bias?type=team)
prepare_ds_data.py contains functions for downloading Dollar Street (https://www.gapminder.org/dollar-street) images and preparing them for Income Bias experiments.
Income Bias.py performs experiments testing bias towards "cheap" or "expensive" items for object topics on Dollar Street.
A report detailing the methods and results can be found here: https://github.com/anitavero/twitter_thumbnail/blob/main/Twitter_Algorithmic_Bias___Report.pdf