City events are getting popular and are attracting a large number of people. This increase needs for methods and tools to provide stakeholders with crowd size information for crowd management purposes. Previous works proposed a large number of methods to count the crowd using different data in various contexts, but no methods proposed using social media images in city events and no datasets exist to evaluate the effectiveness of these methods. In this study we investigate how social media images can be used to estimate the crowd size in city events. We construct a social media dataset, compare the effectiveness of face recognition, object recognition, and cascaded methods for crowd size estimation, and investigate the impact of image characteristics on the performance of selected methods. Results show that object recognition based methods, reach the highest accuracy in estimating the crowd size using social media images in city events. We also found that face recognition and object recognition methods are more suitable to estimate the crowd size for social media images which are taken in parallel view, with selfies covering people in full face and in which the persons in the background have the same distance to the camera. However, cascaded methods are more suitable for images taken from top view with gatherings distributed in gradient. The created social media dataset is essential for selecting image characteristics and evaluating the accuracy of people counting methods in an urban event context.@en

Using social media data, such as Twitter and Instagram, I identify relevant information, develop data models, estimate and analyse crowds’ characteristics in city events, including demographic and city-role composition, Spatial-temporal distribution, sentiment estimation, Points of Interest preferences, word use, crowd size and density estimation, which help crowd managers make better decisions.@en

Large-scale events are becoming more frequent in contemporary cities, increasing the need for novel methods and tools that can provide relevant stakeholders with quantitative and qualitative insights about attendees’ characteristics. In this work, we investigate how social media can be used to provide such insights. First, we screen a set of factors that characterize crowd behavior and introduce a set of proxies derived from social media data. We characterize the crowd in two city-scale events, Sail 2015 and King’s Day 2016, analyzing several properties of their attendees, including demographics, city-role, crowd temporal distribution, social media post locations, Point of Interest (PoI.) preferences, and word use. We show that it is possible to characterize crowds in city-scale events using social media data, thus paving the way for new real-time applications on crowd monitoring and management for city-scale events.@en

City events are being organized more frequently, and with larger crowds, in urban areas. There is an increased need for novel methods and tools that can provide information on the sentiments of crowds as an input for crowd management. Previous work has explored sentiment analysis and a large number of methods have been proposed relating to various contexts. None of them, however, aimed at deriving the sentiments of crowds using social media in city events, and no existing event-based dataset is available for such studies. This paper investigates how social media can be used to estimate the sentiments of crowds in city events. First, some lexicon-based and machine learning-based methods were selected to perform sentiment analyses, then an event-based sentiment annotated dataset was constructed. The performance of the selected methods was trained and tested in an experiment using common and event-based datasets. Results show that the machine learning method LinearSVC achieves the lowest estimation error for sentiment analysis on social media in city events. The proposed event-based dataset is essential for training methods to reduce estimation error in such contexts.@en

City-scale events attract large amounts of attendees in temporarily re-purposed urban environments. In this setting, the real-time measurement of the density of attendees stationing in – or moving through – the event terrain is central to applications such as crowd management, emergency support, and quality of service evaluation. Sensing or communication infrastructures (e.g. sensor networks, mobile phones) can be deployed to estimate the number of attendees currently occupying an area. However, the adoption of these technologies is hindered by their cost or sensing resolution. There is evidence that social media data can provide a real-time and semantically rich insight into attendees’ behaviour during city-scale events. Their suitability as a data source for attendees density estimation is yet to be investigated. With this paper we aim at filling this knowledge gap by studying how micro-posts harvested from social media can be used during city-scale events to estimate the density of attendees in a given terrain. To cope with issues of temporal and spatial resolution, we propose 3 classes of density estimation strategies, i.e. geo-based, speed-based and flow-based strategy, inspired by elements of pedestrian traffic flow theory that were successfully assessed during city-scale events. We study the performance of these strategies in the context of SAIL Amsterdam 2015 (Sail) and Kingsday Amsterdam 2016 (Kingsday), two city-scale events that attracted 2 and 1.5 million of attendees in the span of 5 days and 1 day, respectively. We defined four experimental terrains for the Sail event and one for the Kingsday event, and compare density estimates from social media data with measures obtained from counting systems and Wi-Fi sensors. Results show the potential of solutions embedding elements from pedestrian traffic flow theory, which yielded estimates with strong temporal correlations with the sensor observation, and limited mean errors.

@en