Abstract: Despite the outpouring of social support posted to social media channels in the aftermath of disaster, finding and managing content that can translate into community relief, donations, volunteering, or other recovery support is difficult due to the lack of sufficient annotated data around volunteerism. This paper outlines three experiments to alleviate these difficulties. First, we estimate to what degree volunteerism content from one crisis is transferable to another by evaluating the consistency of language in volunteer-and donation-related social media content across 78 disasters. Second it introduces methods for providing computational support in this emergency support function and developing semi-automated models for classifying volunteer-and donation-related social media content in new disaster events. Results show volunteer-and donation-related social media content is sufficiently similar across disasters and disaster types to warrant transferring models across disasters, and we evaluate simple resampling techniques for tuning these models. We then introduce and evaluate a weak-supervision approach to integrate domain knowledge from emergency response officers with machine learningmodelstoimproveclassification accuracy andacceleratethisemergencysupportinnewevents. This method helps to overcome the scarcity in data that we observe related to volunteer-and donation-related social media content.

Abstract: Radiation measurements are key information in post-nuclear accident situations. Automated Twitter accounts have been used to share the readings, but often in an incomplete way from the perspective of data sharing and risk communication between citizen and radiation experts. In this paper, we investigate the requirements for radiation measurements completeness, by analyzing the perceived usefulness of several metadata items that may go along the measurement itself. We carried out a benchmark of existing uses, and conducted a survey with both experts and lay citizens. We thus produced a set of guidelines regarding the metadata that should be used, and the way to publish it.

Abstract: We report on the design, development, and evaluation of a user labeling framework for social media monitoring by emergency responders. By labeling Twitter user accounts based on behavior and content, this novel approach identifies tweets from accounts belonging to Individuals generating Personalized content and captures information that might otherwise be missed. We evaluate the framework using training data from the 2018 Camp, Woolsey, and Hill fires. Approximately 30% of the Individual-Personalized tweets contain first-hand information, providing a rich stream of content for social media monitoring. Because it can quickly eliminate most redundant tweets, this framework could be a critical first step in an end-to-end information extraction pipeline. It may also generalize more easily for new disaster events since it relies on general user account attributes rather than tweet content. We conclude with next steps for refining and evaluating our framework in near real-time during a disaster response.

Abstract: Social media like Twitter offer not only an unprecedented amount of user-generated content covering developing emergencies but also act as a collector of news produced by heterogeneous sources, including big and small media companies as well as public authorities. However, this volume, velocity, and variety of data constitute the main value and, at the same time, the key challenge to implement and automatic detection and tracking of independent emergency events from the real-time stream of tweets. Leveraging online clustering and considering both textual and geographical features, we propose, implement, and evaluate an algorithm to automatically detect emergency events applying a ‘glocal’ approach, i.e., offering a global coverage while detecting events at local (municipality level) scale.

Abstract: Content shared in microblogs during disasters is expressed in various formats and languages. This diversity makes the information retrieval process more complex and computationally infeasible in real time. To address this, we propose a classification model for the identification of prominent users who are sharing relevant and exclusive information during the disaster. Users who have shared at least one tweet about the disaster are modeled using three kinds of time-sensitive features, including topical, social and geographical features. Then, these users are classified into two classes using a linear Support Vector Machine (SVM) to evaluate them over the extracted features and identify the most prominent ones. The first results using the actual dataset, show that our model has a high accuracy by detecting most of the prominent users. Moreover, we demonstrate that all the proposed features used by our model are indispensable to achieve this high accuracy.

Abstract: Real-time information from microposts like Twitter is useful for applications in the crisis management domain. Currently, that potentially valuable information remains mostly unused by the command staff, mainly because the sheer amount of information cannot be handled efficiently. Sentiment analysis has been shown as an effective tool to detect microposts (such as tweets) that contribute to situational awareness. However, current approaches only focus on two or three emotion classes. But using only tweets with negative emotions for crisis management is not always sufficient. The amount of remaining information is still not manageable or most of the tweets are irrelevant. Thus, a more fine-grained differentiation is needed to identify relevant microposts. In this paper, we show the systematic evaluation of an approach for sentiment analysis on microposts that allows detecting seven emotion classes. A preliminary evaluation of our approach in a crisis related scenario demonstrates the applicability and usefulness.

Abstract: Crowdsourced Geographic Information (CGI) has emerged as a potential source of geographic information in different application domains. Despite the advantages associated with it, this information lacks quality assurance, since it is provided by different people. Therefore, several authors have started investigating different methods to assess the quality of CGI. Some of the existing methods have been summarized in different classification scheme. However, there is not an overview of the methods employed to assess the quality of CGI in the absence of authoritative data. On the basis of a systematic literature review, we found 13 methods that can be employed to this end.

Abstract: Crowdsourcing is an increasingly popular approach for processing data in response to disasters. While volunteer crowdsourcing may suÿce for high-profile disasters, paid crowdsourcing may be necessary to recruit workers for less prominent events. Thus, understanding the impact of payment schemes on worker behavior and motivation may improve outcomes. In this work, we presented workers recruited from Amazon Mechanical Turk with a disaster response task in which they could provide a variable number of image ratings. We paid workers a fixed amount to provide a minimum number of image ratings, allowing them to voluntarily provide more if desired; this allowed us to examine the impact of dierent amounts of required work. We found that requiring no ratings resulted in workers voluntary completing more work, and being more likely to indicate motivation related to interest on a post survey, than when small numbers of ratings were required. This is consistent with the motivational crowding-out eect, even in paid crowdsourcing. We additionally found that providing feedback on progress positively impacted the amount of work done.

Abstract: This paper analyses the collaborative dynamic of community in response to urgent situations. Community emergencies arising from natural or man-induced threats are considered as exogenous events that stimulate community resources to be unified around the response, action, and recovery activities related to the emergency. A structured equation model is derived to depict the actions of the community system. The system is described in terms of its resources including the propensity to trigger community action and collaboration among diverse groups. The community is profiled with respect to its ability to respond. The system defines the trigger mechanisms that are considered to be the drivers of collaborative action. A simulation model is presented to enact the system emergencies, community profiles, and collaborative response. The results develop an improved understanding of conditions that engage community collaborative actions as illustrated by examples from community research in the EnRiCH and the C-Change community research projects.

Abstract: People increasingly use microblogging platforms such as Twitter during natural disasters and emergencies. Research studies have revealed the usefulness of the data available on Twitter for several disaster response tasks. However, making sense of social media data is a challenging task due to several reasons such as limitations of available tools to analyze high-volume and high-velocity data streams. This work presents an extensive multidimensional analysis of textual and multimedia content from millions of tweets shared on Twitter during the three disaster events. Specifically, we employ various Artificial Intelligence techniques from Natural Language Processing and Computer Vision fields, which exploit different machine learning algorithms to process the data generated during the disaster events. Our study reveals the distributions of various types of useful information that can inform crisis managers and responders as well as facilitate the development of future automated systems for disaster management.

Abstract: This study contributes an understanding of how ICTs and varying information sources work together during emergency alerts. It builds on the prior work on campus active shooter events by examining an organization that used a range of ICTs including mobile devices, social media, organizational tools, and news media, to notify their stakeholders about an emergency. The study design used a survey to capture the responses from a random sample of over 1000 stakeholders-students, faculty, and staff-who were notified of an active shooter emergency. The findings from the first three notifications suggest that messages reaching the most stakeholders were (a) sent by official sources through ICTs like mobile phones; (b) official email communication, and (c) messages that included face-to-face communication. While 11 different ICTs were included in the study, mass media (i.e., television and radio), and social media (Twitter and Facebook) did not function substantially in the emergency alert process.

Abstract: Several studies on Twitter usage during disasters analyze tweets collected using ad-hoc keywords pre-defined by researchers. While recent efforts have worked to improve this methodology, open questions remain about which keywords can be used to uncover tweets contributing to situational awareness (SA) and the quality of tweets returned using different terms. Herein we consider a novel methodology for uncovering relevant keywords one can use to search for tweets containing situational awareness. We provide a description of the methodology and initial results which suggest our approach may lead to better keywords to use for keyword searching during disasters.

Abstract: In 2020, we have witnessed an unprecedented crisis event, the COVID-19 pandemic. Various questions arise regarding the nature of this crisis data and the impacts it would have on the existing tools. In this paper, we aim to study whether we can include pandemic-type crisis events with general non-pandemic events and hypothesize that including labeled crisis data from a variety of non-pandemic events will improve classification performance over models trained solely on pandemic events. To test our hypothesis we study the model performance for different models by performing a cross validation test on pandemic only held-out sets for two different types of training sets, one containing only pandemic data and the other a combination of pandemic and non-pandemic crisis data, and comparing the results of the two. Our results approve our hypothesis and give evidence of some crucial information propagation upon inclusion of non-pandemic crisis data to pandemic data.