Common Characteristics Found in Twitter Data and Human Behavior on Retweet

Abstract

In order to see the human behavior on Twitter, we analyzed a large amount of Twitter data collected by the keyword search function of Twitter API. We found some common features in Twitter data; e.g., for any search keywords, the number of retweets follows a power-law distribution. We describe some of these characteristics based on the assumption that Twitter users tend to retweet tweets that have been retweeted many times before.