Storm Blueprints: Patterns for Distributed Real-time Computation

Graph data model

The primary entity in our data model is a Twitter user. A Twitter user can perform the following relationship-forming actions when posting a tweet:

• Use a word
• Mention a hashtag
• Mention another user
• Mention a URL
• Retweet another user

This concept maps very naturally into a graph model. In the model, we will have four different entity types (vertices):

• User: This represents a Twitter user account
• Word: This represents any word contained in a tweet
• URL: This represents any URL contained in a tweet
• Hashtag: This represents any hashtag contained in a tweet

Relationships (edges) will consist of the following actions:

• mentions_user: Using this action, ...