A retweet network models users retweeting posts from other users as a graph. In a labeled-property graph, users publish tweets, and retweets connect users with the tweets they amplify. section 2.3 section 2.3.3
Choose the representation according to the questions the analysis must answer:
- Represent a retweet directly as a RETWEETS relationship between the original tweet and the user who retweeted it. section 2.3.3
- Represent the retweet as a separate tweet node when later analysis may need to attach additional relationships to that retweet. section 2.3.3
- Count retweets to identify popular topics or authors, or infer relationships between users from repeated retweeting. section 2.3.3
- Translate the indirect pattern into a directed, weighted AMPLIFY relationship; its weight equals the number of retweets between users. section 2.3.3
From retweet to amplification
If Vanessa retweets Kim's tweet, the graph can represent that indirect interaction as a direct AMPLIFY relationship from Kim to Vanessa. One retweet gives the relationship weight 1; repeated retweets increase the weight. section 2.3.3
(:User)-[:PUBLISH]->(:Tweet)<-[:RETWEETS]-(:User)
Use this compact pattern to inspect the users, published tweets, and users who retweeted them. section 2.3.3
In Neo4j Browser, sample RETWEETS relationships can be visualized by adapting a generated Cypher query for another relationship type. section 4.1
