Using social media for sub-event detection during disasters

Abstract

Social media platforms have become fundamental tools for sharing information during natural disasters or catastrophic events. This paper presents SEDOM-DD (Sub-Events Detection on sOcial Media During Disasters), a new method that analyzes user posts to discover sub-events that occurred after a disaster (e.g., collapsed buildings, broken gas pipes, floods). SEDOM-DD has been evaluated with datasets of different sizes that contain real posts from social media related to different natural disasters (e.g., earthquakes, floods and hurricanes). Starting from such data, we generated synthetic datasets with different features, such as different percentages of relevant posts and/or geotagged posts. Experiments performed on both real and synthetic datasets showed that SEDOM-DD is able to identify sub-events with high accuracy. For example, with a percentage of relevant posts of 80% and geotagged posts of 15%, our method detects the sub-events and their areas with an accuracy of 85%, revealing the high accuracy and effectiveness of the proposed approach.

E-ware: a big data system for the incremental discovery of spatio-temporal events from microblogs

“…Moreover, the approaches presented in Belcastro et al. ( 2021 ), Dabiri and Heaslip ( 2019 ), Shah and Dunn ( 2019 ) are monitoring specific types of event detection, which cannot be fairly compared against the generic approaches that consider unsupervised learning techniques. Furthermore, Rehman et al.…”

“…The study presented in Belcastro et al. ( 2021 ) is specific-purpose detection technique for discovering sub-events after a disaster, which uses classic classification and DBSCAN clustering. This approach does not take the temporal and incremental features into account and cannot be generalized to all types of events.…”

“…( 2021 ) Generic Pairwise popularity GCN and Heterogeneous DBSCAN for social event classification ✓ ✓ ✓ ✓ Near-real-time Belcastro et al. ( 2021 ) Specific: discovering sub-events after a disaster Classic classification and DBSCAN clustering ✓ ✓ Batch Dabiri and Heaslip ( 2019 ) Specific: traffic event detection CNN+LSTM for classification, word embeddings and geocoding ✓ ✓ Batch George et al. ( 2021 ) Generic Poisson model and quadtree structure for multiscale event detection ✓ ✓ ✓ Batch Han et al.…”

Urgent Computing for Protecting People From Natural Disasters

“…6 For handling these kind of occurrences, we and our colleagues proposed a system called Sub-Events Detection on sOcial Media During Disasters (SEDOM-DD), for detecting and signaling subevents during disasters. 7 Specifically, the designed system addresses two important issues: understanding whether a post is relevant about a disaster, and discovering the subevents that occurred in the disaster area. SEDOM-DD performs these tasks in four main steps: 1) collecting posts that are potentially related to the disaster, 2) filtering posts to keep only the relevant ones, 3) enrichment data by using information contained in posts to increase the number of posts for which it is possible to estimate their geolocalization, and 4) using clustering techniques on geotagged relevant posts for detecting subevents.…”

A Novel Burst Event Detection Model Based on Cross Social Media Influence

“…Social public emergencies such as COVID-19, fake comments and news on social media have garnered the interest and attention of worldwide Internet users [1,2,3,4,5], who comment, retweet, and distribute information about the events on key social media networks platforms such as Twitter [6], Facebook, WhatsApp, Weibo, WeChat, and BBS. [7,8]. This leads to a relatively high degree of speech liberalization on the network because any user, at any time and place, can freely discuss any emergency topic [9,10].…”