Abstract: Semi-structured interviews were conducted with U.S. public sector emergency managers to probe barriers to use of social media and reactions to possible software enhancements to support such use. The three most frequently described barriers were lack of personnel time to work on use of social media, lack of policies and guidelines for its use, and concern about the trustworthiness of pulled data. The most popular of the possible technological enhancements described for Twitter are filtering by category of user/contributor, and display of posts on a GIS system with a map-based display.

Abstract: In summer of 2014 the city of Münster experienced an urban flash flood not seen before with such intensity in Germany. This paper investigates the subsequent governmental and ad-hoc community response actions with a focus on the chronologies of Facebook and Twitter usage. Interviews identified drawbacks of coordinating volunteers in social media ecosystems. Possible solutions to overcome issues related to the interaction of community and official relief activities are identified.

Abstract: Significant progress has been made towards automated classification of disaster or crisis related tweets using machine learning approaches. Deep learning models, such as Convolutional Neural Networks (CNN), domain adaptation approaches based on self-training, and approaches based on pre-trained language models, such as BERT, have been proposed and used independently for disaster tweet classification. In this paper, we propose to combine self-training with CNN and BERT models, respectively, to improve the performance on the task of identifying crisis related tweets in a target disaster where labeled data is assumed to be unavailable, while unlabeled data is available. We evaluate the resulting self-training models on three crisis tweet collections and find that: 1) the pre-trained language model BERTweet is better than the standard BERT model, when fine-tuned for downstream crisis tweets classification; 2) self-training can help improve the performance of the CNN and BERTweet models for larger unlabeled target datasets, but not for smaller datasets.

Abstract: Many machine learning and natural language processing techniques, including supervised and domain adaptation algorithms, have been proposed and studied in the context of filtering crisis tweets. However, applying these approaches in real-time is still challenging because of time-critical requirements of emergency response operations and also diversities and unique characteristics of emergency events. In this paper, we explore the idea of building “generalized” classifiers for filtering crisis tweets that can be pre-trained, and are thus ready to use in real-time, while generalizing well on future disasters/crises data. We propose to achieve this using simple feature based adaptation with tweet representations based on word embeddings and also sentence-level embeddings, representations which do not rely on unlabeled data to achieve domain adaptations and can be easily implemented. Given that there are different types of word/sentence embeddings that are widely used, we propose to compare them to get a general idea about which type works better with crisis tweets classification tasks. Our experimental results show that GloVe embeddings in general work better with the datasets used in our evaluation, and that the supervised algorithms used in our experiments benefit from GloVe embeddings trained specifically on crisis data. Furthermore, our experimental results show that following GloVe, the sentence embeddings have great potential in crisis tweet tasks.

Abstract: This paper describes the design, execution, and results of a participatory design workshop with emergency public information officers (PIOs). During the workshop, PIOs and researchers explored ideas and designs for supporting the social media needs of PIO work. Results indicate that PIO perceptions of social media have changed as they have learned to incorporate activities of the public into their work, yet they still struggle with issues of trust and liability. Based on workshop design activities, the paper offers a set of design recommendations for supporting the social media needs of PIO work practice such as the ability to monitor, document, and report social media activity.

Abstract: Social media has recently become a digital lifeline used to relay information and locate survivors in disaster situations. Currently, officials and volunteers scour social media for any valuable information; however, this approach is implausible as millions of posts are shared by the minute. Our goal is to automate actionable information extraction from social media posts to efficiently direct relief resources. Identifying damage and human casualties allows first responders to efficiently allocate resources and save as many lives as possible. Since social media posts contain text, images and videos, we propose a multimodal deep learning framework to identify damage related information. This framework combines multiple pretrained unimodal convolutional neural networks that extract features from raw text and images independently, before a final classifier labels the posts based on both modalities. Experiments on a home-grown database of labeled social media posts showed promising results and validated the merits of the proposed approach.

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: Microblogging sites such as Twitter can play a vital role in spreading information during “natural” or man-made disasters. But the volume and velocity of tweets posted during crises today tend to be extremely high, making it hard for disaster-affected communities and professional emergency responders to process the information in a timely manner. Furthermore, posts tend to vary highly in terms of their subjects and usefulness; from messages that are entirely off-topic or personal in nature, to messages containing critical information that augments situational awareness. Finding actionable information can accelerate disaster response and alleviate both property and human losses. In this paper, we describe automatic methods for extracting information from microblog posts. Specifically, we focus on extracting valuable “information nuggets”, brief, self-contained information items relevant to disaster response. Our methods leverage machine learning methods for classifying posts and information extraction. Our results, validated over one large disaster-related dataset, reveal that a careful design can yield an effective system, paving the way for more sophisticated data analysis and visualization systems.

Abstract: Integration of social media in emergency response environments presents specific organizational challenges, such as lack of resources or information credibility. Additionally, there exists individual resistance to change in these environments that could potentially discourage adoption. To identify and understand these challenges, we conducted semi-structured group interviews with emergency call takers and dispatchers. We find that these PSAP operators desire participation and explanation of changes throughout the organizational change process. Participants also articulated they desired training regarding change even when not directly affected. Though change management procedures often call for these strategies, they are commonly overlooked, leaving individuals to imagine worse case scenarios that manifest as additional stress in an already stressful work environment. It is suggested that a formalized change management process which directly addresses the identified challenges within the organizational technology adoption process (OTAP) is needed in order to mitigate undue stress.

Abstract: On September 8, 2011 the Great San Diego/Southwest Blackout occurred. Approximately 5 million people were affected by this blackout. This paper explores the availability of social media following such a crisis event. Contrary to expectations, the cell phone system did not have the expected availability and as a result, users had a difficult time using social media to status/contact family and friends. This paper presents a survey exploring the use and availability of social media during the Great San Diego/Southwest Blackout event. © 2012 ISCRAM.

Abstract: Social media enables fast information exchange and status reporting during crises. Filtering is usually required to
identify the small fraction of social media stream data related to events. Since deep learning has recently shown to
be a reliable approach for filtering and analyzing Twitter messages, a Convolutional Neural Network is examined for
filtering crisis-related tweets in this work. The goal is to understand how to obtain accurate and robust filtering
models and how model accuracies tend to behave in case of new events. In contrast to other works, the application
to real data streams is also investigated. Motivated by the observation that machine learning model accuracies
highly depend on the used data, a new comprehensive and balanced compilation of existing data sets is proposed.
Experimental results with this data set provide valuable insights. Preliminary results from filtering a data stream
recorded during hurricane Florence in September 2018 confirm our results.

Abstract: Overload reduction is essential to exploit Twitter text data for crisis management. Often used pre-trained machine learning models require training data for both, crisis-related and off-topic content. However, this task can also be formulated as a one-class classification problem in which labeled off-topic samples are not required. Gaussian processes (GPs) have great potential in both, binary and one-class settings and are therefore investigated in this work. Deep kernel learning combines the representative power of text embeddings with the Bayesian formalism of GPs. Motivated by this, we investigate the potential of deep kernel models for the task of classifying crisis-related tweet texts with special emphasis on cross-event applications. Compared to standard binary neural networks, first experiments with one-class GP models reveal a great potential for realistic scenarios, offering a fast and flexible approach for interactive model training without requiring off-topic training samples and comprehensive expert knowledge (only two model parameters involved).

Abstract: Crisis informatics researchers have taken great interest in methods to identify information relevant to crisis events posted by digital bystanders on social media. This work codifies the information needs of emergency dispatchers and first responders as a method to identify actionable information on social media. Through a design workshop with public safety professionals at a Public-Safety Answering Point (PSAP) in the United States, we develop a set of information requirements that must be satisfied to dispatch first responders and meet their immediate situational awareness needs. We then present a manual coding scheme to identify information satisfying these requirements in social media posts and apply this scheme to fictitious tweets professionals propose as actionable information to better assess ways that this information may be communicated. Finally, we propose automated methods from previous literature in the field that can be used to implement these methods in the future.