Data Analytics and Visualization for Social Media

The exponential increase in social media textual data creates enormous challenges to read and interpret text.
Twitter has grown from delivering 65 million tweets per day to over 200 million in 1 year. Facebook has 800
million users (each with ~130 friends). Over half the users visit Facebook daily. Of active Internet users, 77%
read blogs. These rapidly growing forms of communication have society struggling to understand and exploit.
Online social media (Twitter, Facebook …) allow readers to express thought/opinions on content. Media
publications have added opinion blog commentary to provide readers’ opportunity to share comments.
Editorial/business leaders see value in audience communication and appreciate that learning aspects such as
emotional tone influence their offerings, contributors and readers. Discovering emotion in text can positively
impact sales, investment, and provide a deeper knowledge of the influence different authors have on the public.
We will develop methods and tools that discern meaning from social media text. We will identify the
emotional content within text and develop tools to better understand emotions found in a social media text, and
develop new tools for media users to shape emotional content and respond to others.
Application of such methods and tools are limitless. Business relies on direct consumer dialogue through
social media to engender loyalty and predict and understand consumer behaviors. In the media, editorial
leadership is eager to better manage reader commentary and to learn underlying patterns that suggest specific
emotional tone. In healthcare, free-form texts are the most valuable data (doctor’s notes, patient histories,
healthcare messages posted by patients on social media). Such text contain information for physicians to use in
practice and for public agencies to make healthcare decisions. Methods and tools that discern meaning by
extracting information piecewise and visualize complex data represent an important advance in our
understanding of social media intercourse. A data-driven design approach to visualize content will aggregate
meanings more apparently and improve our ability to understand this emerging communication channel.

Faculty Supervisor:

Nick Cercone


An Than



Engineering - computer / electrical



York University



Current openings

Find the perfect opportunity to put your academic skills and knowledge into practice!

Find Projects