Emojis are very popular among online users. People use them in short texts they send or post because they are useful in conveying users' feelings. Despite their wide use, we still do not understand the patterns in their usage. In this paper, we present an important first step to-wards this understanding by using a modeling based on Network Science. We create a co-occurrence bipartite emoji-word network from 6 collections of tweets; each from a different topic of conversation. We then present results regarding the sentiment of an emoji as well as its semantics; its connection to words that define its meaning. Our results show that emojis are generally used in a positive sentiment but the semantics differ depending on the subject of the conversation.
展开▼