Given Twitter and Instagram posts, can we estimate the density of crowds attending city events, such as national holiday festival or parade? In this study, we propose three density estimation strategies (i.e. geo-based, speed-based, and flow-based) and compare their performance with counting systems and Wi-Fi sensors in two city events. [Link]
Can tweets or Instagram posts reflect the sentiment of people in a festival? An Interesting and critical challenge for crowd manager in city events. In this study, we explore a set of methods, e.g. machine learning methods and lexicon-based methods, to estimate sentiment of crowd from their social media posts. [Link]
Can tweets or Instagram posts characterise crowds in a festival, in terms of demographic and such? Using social media data collected during a national holiday in a city, is it possible to characterise the attendees of this event in terms of their age, gender, and what did they talk about? In this study, we use explore effective methods to derive a set of information about crowds in such context. [Link]
Estimate number of people in the crowds is essential for density estimation. On Instagram, people share selfie pictures, group pictures or panorama views during attending city events. Knowing the number of people showing in the Instagram is useful for density estimation, and on top of that, estimate the crowded level during the events. In this work, we investigate state-of-the-art methods for estimating the number of people in social media images during city events. [Link]