Trump is Kim Jong Un With Worse Hair by Roopa Vasudevan

Social media data relating to politics and political candidates is often considered to be a fertile alternative to traditional polling practices, the drawbacks of which have become more apparent in recent US election cycles. However, the data is often forced into numerical structures, stripped of character and nuance, and graphed to reflect garden-variety approval and disapproval numbers, when in reality, public opinion -- as expressed both on and offline -- is much more varied and mercurial. Is it possible to harness this fickle and multifaceted nature, and use it for our benefit? And, should we be looking at social media in its own right, rather than just as a replacement for polls?

In this artist talk and workshop, Roopa will first review her artwork and research surrounding American polling practices, social media, and why representing public opinion as a strict dichotomy isn't always the best route. Then, we'll experiment with some qualitative analysis of our own, using Twitter data surrounding the current Presidential administration, as well as basic text processing tools like regular expressions and part of speech tagging. Please come with a version of Python installed on your laptop (the demo will be in Python 3, although 2 works as well), and with the TextBlob module installed -- instructions available here.

Roopa Vasudevan is an American artist, creative coder, and researcher, currently based in New York, NY and Shanghai, China. Her work has been exhibited internationally and featured on Reuters, PBS NewsHour, Slate, Hyperallergic, Jezebel, Complex, PSFK, the FADER, Public Radio International, and more, as well as on American, French and German television. She is a member of the Flux Factory artist community (based in Queens, NY) and is currently an Assistant Arts Professor of Interactive Media Arts at NYU Shanghai.