Tweet Emotion Mapping: Understanding US Emotions in Time and Space

Twitter is one of the most popular social media platform where users post their views and emotions on a regular basis. Consequently, Twitter tweets have become a valuable knowledge source for emotion analysis. In this paper, we present a new framework for tweet emotion mapping and emotion change analysis. It introduces a novel, generic spatio-temporal data analysis and storytelling framework and its architecture. The input for our approach are the location and time were and when the tweets were posted and an emotion assessment score in [-1, +1], with +1 denoting a very positive emotion and -1 a very negative emotion. Our first step is to segment the input dataset into batches with each batch containing tweets that occur in a specific time interval, for example weekly, monthly or daily. Next, by generalizing existing kernel density estimation techniques, we transform each batch into a continuous function that takes positive and negative values. Next, we use contouring algorithms to find contiguous regions with highly positive and highly negative emotions for each of the batch. After that, we apply a generic, change analysis framework that monitors how positive and negative emotion regions evolve over time. In particularly, using this framework unary and binary change predicate are defined and matched against the identified spatial clusters, and change relationships will then be recorded, for those spatial clusters for which a match occurred. Finally, we propose animation techniques to facilitate spatio-temporal data storytelling based on the obtained spatio-temporal data analysis results. We demo our approach using tweets collected in the state of New York in June 2014.