HeyLo: Visualizing User Interests from Twitter Using Emoji in Mixed Reality
Additional Funding Sources
This project is supported by a 2019-2020 STEM Undergraduate Research Grant from the Higher Education Research Council.
Abstract
We tackle the problem of analyzing a user's interests from social media content and subsequently visualizing these interests in an extended reality environment. We compare five models for extracting interests from Twitter users and how we can measure the effectiveness of these models. We also look at how these interest extraction models fit in the context of HeyLo, an extended reality computational creativity (XRCC) framework for visualizing potential conversational topics. The chosen interests for a particular person are visualized using emoji. We accomplish this by using an emoji2vector model to find the closest related emoji to a given interest. We perform a comparative analysis between the five interest extraction models on real-world users and their tweets, evaluating specificity, variance, and relevance.
HeyLo: Visualizing User Interests from Twitter Using Emoji in Mixed Reality
We tackle the problem of analyzing a user's interests from social media content and subsequently visualizing these interests in an extended reality environment. We compare five models for extracting interests from Twitter users and how we can measure the effectiveness of these models. We also look at how these interest extraction models fit in the context of HeyLo, an extended reality computational creativity (XRCC) framework for visualizing potential conversational topics. The chosen interests for a particular person are visualized using emoji. We accomplish this by using an emoji2vector model to find the closest related emoji to a given interest. We perform a comparative analysis between the five interest extraction models on real-world users and their tweets, evaluating specificity, variance, and relevance.