| Spring 2020
Meet Engelbart, a real-time web application that presents Twitter in an agent-based environment. Completed with Garrett Vercoe and advised by Ehsan Baharlou. Research paper presented and published online at ACADIA 2020.
With the goal of popping the social media filter bubble, Engelbart reframes the discussions on Twitter by presenting them entirely differently than the algorithmically ordered feed.Engelbart is the interactive agent that explores this space.Topic clusters form the structure of the space, with more related topics positioned closer to each other. The activity level of discussions for each topic is also reflected in the size of the cluster.The color of activity within the clusters is determined by sentiment analysis on aggregated data within the topic. More green in a cluster means that discussion is generally positive!Engelbart makes his way around the space, visiting topic clusters and revealing information about them as he learns what more about them himself.Behind the scenes, many layers of data analysis in the form of natural language processing are condensed to deliver information in a concise and honest way.