This study aimed to delve into the discourse about contraception on Reddit, given that it's a taboo topic in this conservative country, and to understand the sentiment of Filipinos on the topic. In creating the dataset, PRAW (Python Reddit API Wrapper) was utilized for gathering submissions whose titles and self-texts include certain keywords related to contraception. To analyze the comment and submission texts, TFIDF and N-gram vectorizers, Wordcloud, VADER sentiment analysis, and LDA were used. A neural network using an LSTM model was used to classify the stances of the comments based on whether they support, deny, question, or comment on their parent comments to further understand the discourse. We conclude that many Filipino Redditors support the use of contraception based on their stance, sentiment and general word usage and topic discussion in the online community. Vectorizing the parent and child comments as features might not also be sufficient for classifying stance, based on the LSTM model.

You can find the link to our paper here: https://github.com/daniddelrio/Philippine-Contraception-Discourse/blob/master/mr_final_delrio_lim.pdf