Weighted Lexicon-based Sentiment Analysis for Women Career Traits in Information Technology
Kaitlin De Chastelain Finnigan, Fahim Anzum, Jon George Rokne, Marina L. Gavrilova · 2022
In this paper, an unsupervised sentiment analysis model leveraging the Polarity Rank algorithm and sentence dependency graph is proposed to predict the sentiments of women in IT. The primary objective is to identify groups of individuals with common traits (community and job role) who had different sentiments, as well as explore the impact of Diversity Equity & Inclusion (DEI) programs on participants' career satisfaction. The model, based on a 7-point Likert scale, classified responses into Very Negative, Negative, Slightly Negative, Neutral, Slightly Positive, Positive, and Very Positive classes By delving into different demographics and different questions within the survey, it was found that DEI programs have a positive impact on career satisfaction as well as lessening the presence of Anger, Fear, and Sadness. Additionally, it was discovered that Anticipation was a dominant emotion in all responses. This paper provides a foundational look at the sentiments expressed by women in the IT industry.