Sentiment Analysis of Participants Interactions in a Hackathon Context: The Example of a Slack Corpus

This paper presents the analysis of participants' interactions during an online hackathon using Natural Language Processing (NLP) techniques. In particular, we explored the communication of groups facilitated by Slack focusing on the use of emojis. Our findings suggest that most used emojis are positive, while negative emojis appeared rarely. Sentiment of written messages was overall positive and could be linked to topics such as motivation or achievements. Topics about participants' disappointment regarding their progress or the hackathon organization, technical issues and criticism were associated with negative sentiment. We envision that our work offers insights regarding online communication in group and collaborative contexts with an emphasis on group work and interest-based activities.