Citation Context

...ensor (Aiello et al., 2013) system facilitates the tracking of predefined events using social streams. In contrast, we track all automatically discovered events we find in the stream. The Twitcident (=-=Abel et al., 2012-=-) project deals with userdriven searching through Social Media with respect to crisis management. However, unlike ReDites, these crises are not automatically discovered. The LRA Crisis Tracker2 has a ...

Citation Context

... have considered scaling to the Twitter fire hose. Research has also addressed event monitoring but mostly in news stories (e.g., [7]). Our event monitoring work is perhaps most similar to Twitcident =-=[3]-=-, but that work uses deep semantic techniques (based on named entity recognition) that tend to be error prone, and hence achieves limited accuracy. Regarding scalable stream processing, recent works (...

Citation Context

...ace allows rescue organizations to parametrize emergency filters, visualize emergency information on the map and summarize the content of emergency messages through tag clouds. The Twitcident project =-=[5]-=- goes one step further and enriches structured emergency information with data obtained from Twitter streams. They use natural language processing (NLP) techniques, more specifically part-of-speech (P...

Citation Context

...s arranged opposite each other and one joint workspace in the middle (Figure 3). The design for two persons ensures that each user has its own workspace, which is big enough for independent work, and there is still sufficient space for the joint workspace. In the following, we will present the functionality of the application CoTable which we have implemented. Figure 3: Overview of CoTable A: The individual workspace. Initially, there is a single workspace at the long side of the table (Figure 4). This workspace comprises (1) a function for searching keywords or locations within social media, (2) a list view of the search results, and (3) the option to activate the second workspace. The search for keywords or locations is implemented via a REST-interface and enables the continuous extraction of contents from social media, which correspond to the location as well as the keyword and its synonyms. To handle information overload for emergency services, CoTable is just a small part of an architecture and the content provided by this API is already pre-filtered by volunteers (Reuter et al., 2015). This pre-filtering ensures a summarization of relevant A B C D E Ludwig et al. CoTable: Social...

Citation Context

...he continuous flow of tweets, posts and announcements on websites manually. In the Netherlands alone, millions of tweets are posted each day. Therefore, it is important to have an automatic system that automatically selects and ranks threatening messages, enabling security operators to check the relevant messages much more efficiently. To date, little research has been published about recognizing death threats on social media. There are several approaches to perform OSINT for disaster management [1], incident detection by handcrafted rules and semantic filtering using DBpedia or news agencies [2], to handle limited availability of annotated data with transfer or semi-supervised learning [3], [4], but they are not optimized for the detection of threats by individuals and they do not exploit the relations between words in a message. Two approaches for recognizing threats in the Dutch language have been reported [5], [6], which are both based on N-grams. However, their performance is poor on large test sets. Very recently, [6] was improved with encouraging results [7]. However, a side-by-side comparison with our work is not straightforward since the used datasets are different. In this p...

Citation Context

...l be validated through this study, and it is hoped that the methodology developed can be extrapolated to other urban environments as well, with a second validation study already planned to take place next year in London, United Kingdom. 1. Introduction In recent years automated processing of rich, geo-tagged, social media text streams, such as Tweets and Facebook status updates is receiving considerable attention in the literature. This is largely motivated by the insights and value that such datasets were shown to provide (Chew & Eysenbach, 2010; O’Connor et al., 2010; Tumasjan et al., 2010; Abel et al., 2012). Social-media streams, in general, alStresscapes: Validating Linkages between Place and Stress Expression on Social Media low for observing large numbers of spontaneous, real-time interactions and varied expression of opinion, which are often fleeting and private (Miller, 2011). Miller (2011) further points out that some social scientists now see an unprecedented opportunity to study human communication with various applications and contexts, which has been an obstacle up until recently. O’Connor et al. (2010) demonstrated how large-scale trends can be captured from Twitter messages, based on...

Citation Context

...]. an initial query for collecting potentially relevant Twitter messages (tweets) using the Twitter Streaming API 7 . Four steps as shown in Fig. 1.6 are applied for improving the filtering of tweets =-=[8]-=-. First, based on the incident information, an incident profile is created which describes the attributes of the incidents. This might be related concepts like locations or persons and weights of thei...