Disaster Impacts Surveillance from Social Media with Topic Modeling and Feature Extraction: Case of Hurricane Harvey

Abstract

Twitter can supply useful information on infrastructure impacts to the emergency managers during major disasters, but it is time consuming to filter through many irrelevant tweets. Previous studies have identified the types of messages that can be found on social media during disasters, but few solutions have been proposed to efficiently extract useful ones. We present a framework that can be applied in a timely manner to provide disaster impact information sourced from social media. The framework is tested on a well-studied and data-rich case of Hurricane Harvey. The procedures consist of filtering the raw Twitter data based on keywords, location, and tweet attributes, and then applying the latent Dirichlet allocation (LDA) to separate the tweets from the disaster affected area into categories (topics) useful to emergency managers. The LDA revealed that out of 24 topics found in the data, nine were directly related to disaster impacts—for example, outages, closures, flooded roads, and damaged infrastructure. Features such as frequent hashtags, mentions, URLs, and useful images were then extracted and analyzed. The relevant tweets, along with useful images, were correlated at the county level with flood depth, distributed disaster aid (damage), and population density. Significant correlations were found between the nine relevant topics and population density but not flood depth and damage, suggesting that more research into the suitability of social media data for disaster impacts modeling is needed. The results from this study provide baseline information for such efforts in the future.

Reducing Social Media Attention Inequality in Disasters: The Role of Official Media During Rainstorm Disasters in China

“…We also confirmed the baseline results with a panel regression model over a long time window (Xiao et al 2015;Robertson and Feick 2016). Most of the existing literature has analyzed social media attention inequality for individual disasters (Takahashi et al 2015;Cheng et al 2019;Mihunov et al 2022). However, it is difficult to explore whether social media attention inequality persists after local infrastructure and awareness of disaster relief have improved (Zou et al 2019;Dargin et al 2021).…”

“…Analysis of social media (Facebook, Twitter, Weibo, and so on) data has proven to be an efficient method for evaluating disasters and lowering disaster risk (Ogie et al 2019). Millions of people turn to social media during both natural hazard and human-induced disasters (Mihunov et al 2022). Humanitarian organizations, government agencies, and public health authorities are progressively utilizing realtime data from social media to assess disaster risks, save lives, and provide assistance to those in distress (Vieweg 2012).…”

“…This becomes particularly crucial when traditional information sources such as television or radio are unavailable (Castillo 2016). During disasters, social media can be used to study the extent of the disaster-related damage (Ogie et al 2019;Mihunov et al 2022), disseminate information (Ragini et al 2018;Karami et al 2020), coordinate the distribution of aid (Fan et al 2020), examine the mental health of affected individuals (Ahmed and Sinnappan 2013;Kim and Madison 2020;Palen et al 2020;Willson et al 2021), raise money for reconstruction (Ahmed and Sinnappan 2013;Takahashi et al 2015), and more.…”

Positive Determinism of Twitter Usage Development in Crisis Communication: Rescue and Relief Efforts after the 6 February 2023 Earthquake in Türkiye as a Case Study

“…The study by Mihunov et al (2022) explained the use of social media and the impact of deploying such technologies in high-risk scenarios, in this case, Hurricane Harvey, which occurred in the United States in 2017. The researchers wanted to collect user-generated data and information in order to enhance disaster surveillance operations.…”

Social media and volunteer rescue requests prediction with random forest and algorithm bias detection: a case of Hurricane Harvey

“…Hurricane Harvey events left a digital footprint that prompted extensive research by disaster management scholars (Zou et al 2018, Mihunov et al 2022. Availability of social media data and growing accessibility of ML methods make it likely that black-box models predicting locations of rescue request will emerge and be adopted for real-world applications.…”