Characteristics of social media stories

Abstract

An emerging trend in social media is for users to create and publish “stories”, or curated lists of web resources with the purpose of creating a particular narrative of interest to the user. While some stories on the web are automatically generated, such as Facebook’s “Year in Review”, one of the most popular storytelling services is “Storify”, which provides users with curation tools to select, arrange, and annotate stories with content from social media and the web at large. We would like to use tools like Storify to present automatically created summaries of archival collections. To support automatic story creation, we need to better understand as a baseline the structural characteristics of popular (i.e., receiving the most views) human-generated stories. We investigated 14,568 stories from Storify, comprising 1,251,160 individual resources, and found that popular stories (i.e., top 25% of views normalized by time available on the web) have the following characteristics: 2/28/1950 elements (min/median/max), a median of 12 multimedia resources (e.g., images, video), 38% receive continuing edits, and 11% of the elements are missing from the live web.