Traffic Impact Area Detection and Spatiotemporal Influence Assessment for Disaster Reduction Based on Social Media: A Case Study of the 2018 Beijing Rainstorm

Abstract

The abnormal change in the global climate has increased the chance of urban rainstorm disasters, which greatly threatens people’s daily lives, especially public travel. Timely and effective disaster data sources and analysis methods are essential for disaster reduction. With the popularity of mobile devices and the development of network facilities, social media has attracted widespread attention as a new source of disaster data. The characteristics of rich disaster information, near real-time transmission channels, and low-cost data production have been favored by many researchers. These researchers have used different methods to study disaster reduction based on the different dimensions of information contained in social media, including time, location and content. However, current research is not sufficient and rarely combines specific road condition information with public emotional information to detect traffic impact areas and assess the spatiotemporal influence of these areas. Thus, in this paper, we used various methods, including natural language processing and deep learning, to extract the fine-grained road condition information and public emotional information contained in social media text to comprehensively detect and analyze traffic impact areas during a rainstorm disaster. Furthermore, we proposed a model to evaluate the spatiotemporal influence of these detected traffic impact areas. The heavy rainstorm event in Beijing, China, in 2018 was selected as a case study to verify the validity of the disaster reduction method proposed in this paper.

Frontiers of policy and governance research in a smart city and artificial intelligence: an advanced review based on natural language processing

“…This research area has primarily concentrated on enhancing the accuracy and scalability of sentiment analysis techniques, including developing deep learning-based models for more precise classification of the sentiment (Ghahramani et al, 2021;Song et al, 2021;Dutt et al, 2023). Some scholars have also employed NLP to devise algorithms for real-time event detection, such as accidents, natural disasters, and public gatherings (Yang et al, 2020;Zhang et al, 2021a,b), with a focus on refining the integration of event detection systems with other city services for more effective governance.…”

Extracting Disaster-Related Location Information through Social Media to Assist Remote Sensing for Disaster Analysis: The Case of the Flood Disaster in the Yangtze River Basin in China in 2020

“…Compared with some existing studies, including flood disaster assessments based solely on social media [45,46] or remote sensing [47,48], and flood disaster analysis com- In addition, the edges between nodes described the spatial distribution characteristics of people who were concerned about those affected areas. Combined with the corresponding social media data, we can understand why people paid attention to these affected areas and even what requirements people wanted.…”

“…Compared with some existing studies, including flood disaster assessments based solely on social media [45,46] or remote sensing [47,48], and flood disaster analysis combined with multi-source data such as that shown in the literature [14][15][16][17], the method in this paper fully considered disaster-related location information contained in social media texts, constructing the relationship between them and uploading location tags of social media. This not only improves the fusion efficiency of the two kinds of data but also effectively integrates the respective advantages of multi-source data.…”

Road-Related Information Mining from Social Media Data: A Joint Relation Extraction and Entity Recognition Approach

“…Worse is that geotagged posts constitute only a tiny part ranging from 0.42% to 3% of all pieces of social media data [26]. Researchers also try to extract the location information from the social media texts (Tables 1 and 2); they use various predefined dictionaries of nations, states, counties, cities, highways, or roads to recognize the locations [6,12,13,15,16,20,27,28]. This category of lexicon-based methods possesses high precision, but a low recall rate is inflicted when there are OOV locations or any inconsistency (e.g., spelling errors) with predefined terminology lists [29].…”