User-Based Information Search across Multiple Social Media

Abstract

Most of todays Internet users are registered to one or more social media applications. As so many are registered to multiple application, it has become difficult to locate friends, former colleagues, peers and acquaintances. Reasons for this include private profiles, name collisions, multiple usernames, lack of profile attributes and profile picture.

The system designed and implemented in this thesis enable automatic user-based information search across multiple social media without relying on explicit user account registration as a basis for combining information. It uses information from a posting on a social media site to find a specific user on multiple other social media. Twitter is used as a starting point where the end-user can choose a tweet of interest, press the tweet and automatically be presented with information about the author of the tweet, including profiles on other social media.

The application was tested and evaluated by running tweets from a set of 100 Twitter users, where 14 of these were public figures, through the system and documenting the results.