The Engage Corpus: A Social Media Dataset for Text-Based Recommender Systems

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 1885–1889 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 1885 The Engage Corpus: A Social Media Dataset for Text-Based Recommender Systems Daniel Cheng,1, Kyle Yan,1 Phillip Keung,1 Noah A. Smith1,2 1University of Washington, 2Allen Institute for Artificial Intelligence d0@uw.edu, kyleyan@uw.edu, pkeung@uw.edu, nasmith@cs.washington.edu Abstract Social media platforms play an increasingly important role as forums for public discourse. Many platforms use recommendation algorithms that funnel users to online groups with the goal of maximizing user engagement, which many commentators have pointed to as a source of polarization and misinformation. Understanding the role of NLP in recommender systems is an interesting research area, given the role that social media has played in world events. However, there are few standardized resources which researchers can use to build models that predict engagement with online groups on social media; each research group constructs datasets from scratch without releasing their version for reuse. In this work, we present a dataset drawn from posts and comments on the online message board Reddit. We develop baseline models for recommending subreddits to users, given the user’s post and comment history. We also study the behavior of our rec- ommender models on subreddits that were banned in June 2020 as part of Reddit’s efforts to stop the dissemination of hate speech. Keywords: Reddit, social media, recommender systems 1. Introduction User engagement with online communities has come under significant scrutiny in light of the increasing po- larization in global politics. Many researchers have examined social media’s role in recent headlines, like the spread of coronavirus-related vaccine misinforma- tion and the dissemination of conspiracy theories in political discourse (Puri et al., 2020; Douglas et al., 2019, inter alia). Since social media platforms typi- cally use recommender systems to help users discover new content and online groups, a natural question that arises is what role such algorithms have played in chang- ing user behavior. Previous studies and investigations (Horwitz, 2021; Bakshy et al., 2015) have discussed the impact of recommendation algorithms on how users discover and join online groups, and researchers are also interested in understanding how NLP is used in gener- ating recommendations. However, we are not aware of any standardized datasets specifically designed for constructing text-based recommender systems in the so- cial media setting; research groups tend to scrape their own datasets from various web resources without releas- ing their corpora, which hinders the reproducibility of reported results. We introduce the Engage corpus, a large-scale, realis- tic dataset derived from Reddit to address this resource gap. Reddit is a rich source of social media data on a variety of subjects, and there are public APIs that allow researchers to download years of posts and comments for each user. Reddit is divided into 2.5 million com- munities called ‘subreddits’, and users interact within subreddits by writing ‘posts’ or submitting ‘comments’ on existing posts. We say that a user engages with a subreddit if they post …

Content truncated to 500 words