Capturing Racial & Gender Inequities on Social Media Platforms using Machine Learning

Abstract

Online social media platforms provide a continuously evolving database due to the highly increasing popularity and rapid expansion of its user base. Users share their life experiences towards various inequity incidents faced at the workplace on the basis of their race or gender on these platforms while maintaining their anonymity. We aim at utilising famous social media platforms to perform extensive analysis and classification tasks for posts capturing instances of various types of Inequalities prevalent in today’s workplace. We present a framework to mine opinions expressed towards sexual harassment, mental health, racial injustice and gender-based bias in the corporate workplace using NLP techniques on social media data. The documents are represented by semantic similarity to aspect embedding’s captured using an attention-based framework for aspect extraction. In addition, we used scores from Empath categories to add information related to emotional facets.