Architecture of A Model For Emoji Sequences Analysis

With the appearance and widespread use of emojis in messages on social networks, several authors have discussed the possibility of a syntax associated with emojis. In this domain, some authors argue that it is possible to write sentences only with emojis; others consider emojis to be just digital gestures without an associated grammar. In this study, we propose a tool that analyzes co-occurrences of emojis in Twitter messages and the semantic context in which these sequences are used. The approach is to understand if there is meaning in those co-occurrences, if this varies depending on the discussion topic, and if it changes over time. Co-occurrences of emojis will be stored in a graph database. From the text of the tweets, keyphrases will be extracted and stored, with the tweets, in a database in Elasticsearch. The tweets to analyze are related to Brexit, the US elections, and Covid. We will determine the prominent emojis, the most used co-occurrences, with and without emoji repetition, and the associated keyphrases for each discussion topic. This study aims to understand if there is meaning in co-occurrences of emojis - sequences of two emojis - in messages exchanged on social networks as a starting point for a future study on syntax in emojis.