Analysis and Visualization of the Sense Of Place from Twitter Data

Abstract

Social networks (e.g., Twitter or Facebook) are part of our daily life, where we disclose opinions, activities, interests, and so forth. For this reasons, scientific community started to consider people as social sensor in many fields. In this context, we propose a study on the Sense of Place, i.e. the sentiment associated to a geographical location, defined by the geographer Tim Cresswell. We started by collecting and annotating tweets on 4 cities that have English as their mother-tongue, in order to find those which express the Sense of Place. Then, we defined a pipeline to discover both the sentiment (positive or negative) expressed towards the place and the geographical location of the tweet according to its text. The data extracted by the pipeline are then aggregated and visualized by a graphic interface, which allows the exploration of the dataset.