Temporal network analysis of inter-organizational communications on social media during disasters: A study of Hurricane Harvey in Houston
Introduction
Social media (such as Twitter and Facebook) have become a new form of infrastructure for communities in coping with disasters [1,2]. Social media enable timely communication and understanding of situational information. Also, social media offer a participatory, collaborative, and self-organized structure for emergency response and public warning during disasters [[3], [4], [5], [6]]. Accordingly, mapping, modeling and analyzing social media data could provide insights into the network dynamics of inter-organizational communications prior, during, and after a disaster. Understanding the inter-organizational communication networks on online social media is an important step for improving collective action related to hazard communication and public warning among various organizations. Aldrich and Meyer [7] defined links between organizations and links between citizens and organizations as important parts of social capital that have a great contribution to community resilience. Organizations (including government and non-government organizations) conduct and execute emergency responses, and their network interactions would have great impacts on disaster response and post-disaster recovery [[8], [9], [10]]. Recognizing this, many studies have examined inter-organizational communication and coordination during emergency response. For example, Harris [11] studied inter-organizational relationships during and after superstorm Sandy. The analysis identified emergent organizations that connected separate parts of the local stakeholder network, which plays an important role in emergency responses and post-Sandy recovery. Kapucu [12] studied inter-organizational coordination in emergency response to the September 11, 2001, terrorist attacks. The results showed that a well-coordinated inter-organizational network would greatly contribute to effective emergency responses and disaster recovery. These studies mapped and analyzed inter-organizational networks through the use of conventional methods such as surveys. The network models developed through organizational surveys have a major challenge regarding the inclusion of all relevant entities and communication links. With the increasing adoption of social media for communication in emergency response and public warning, there is an important need for characterizing the dynamics of inter-organizational communication networks on social media. However, as discussed in the following section, the existing literature related to social media in civil infrastructure and disaster research offers limited insight into the communication patterns and network properties among organizations during the course of a disaster. Hence, in this paper, we established a methodology to map, model and analyze the inter-organizational communication networks based on data from Twitter to evaluate three fundamental network properties (e.g., degree centrality, betweenness centrality, and core-periphery structure). The study utilized the Twitter data from the Houston area before, during and after Hurricane Harvey to study the evolution of communication patterns, network properties, and roles of organizations at different stages of the disaster.
Section snippets
Background and point of departure
The literature related to social media in disaster research is rapidly growing. Multiple studies, such as Zhang et al. [13]; have provided systematic literature reviews to highlight the state of the field and future directions. In this section, we briefly discuss some recent studies related to analyzing social media in civil infrastructure and disaster research to highlight the point of departure for this study. Twitter, in particular, has been studied extensively by researchers due to its ease
Data
We collected the data related to Hurricane Harvey through the Twitter PowerTrack application programming interface (API) from Gnip, a data provider. The data include the text of the tweets, posted time stamps, retweet/reply status, URLs to external sources, as well as user profile information such as user profession, locality, created time and follower count. To have a focus on the communication among organizations in a specific area, we defined a geographical boundary for geotagged tweets and
Methodology
To extract organizational accounts, map and analyze the inter-organizational communication networks on social media, we propose a three-step framework: (1) extract organization accounts in the dataset, (2) map online organizational communication networks based on twitter interactions among organizational accounts, and (3) conduct network analysis to get daily network properties including in-degree and out-degree centrality, betweenness centrality, and core-periphery structure. Fig. 2
Results
Based on the framework described in the previous section, we mapped the daily online organizational communication networks from August 22 to September 10 (some examples are illustrated in Fig. 3(a)~(f)). Fig. 3(g) shows the number of edges in correspondent daily networks. The weight of edges is the number of interactions between two organizational users on the same day. For example, if one organizational user retweets the post of another organizational user twice on the same day, the weight of
Discussion
The network properties of mapped online organizational communication networks reveal the situational roles of organizational users in online social media during disasters. The following discussion will focus on the evolution of organization roles before, during and after disasters, and the differences between the roles of different types of organizations. Based on the lessons learned from the results, we provide recommendations for organizational users to get better involvement and
Limitation
This study has some limitations. First, despite the best effort of the authors, the dataset might miss some important organizational users that played important roles in Hurricane Harvey such as USACE, Harris County Flood Control District (HCFCD), and FEMA. These organizational users were not included in the dataset because they do not have geographic information mentioned in the Houston area in their twitter biographies (e.g., HCFCD) or they are not physically located in the Houston area
Concluding remarks
This paper mapped and analyzed online organizational communication networks based on interactions among organizational users on Twitter to specify their roles during Hurricane Harvey. The study examined the communication networks and analyzed three network properties to study the roles of organizations in disasters. The findings provide empirical evidence regarding the extent to which different organizational users play a role in social media communications during disasters. With additional
Declaration of competing interest
The authors claim that there is no conflict of interest.
Acknowledgments
The authors would like to acknowledge funding support from the National Science Foundation CRISP project # (1832662): “Anatomy of Coupled Human-Infrastructure Systems Resilience to Urban Flooding: Integrated Assessment of Social, Institutional, and Physical Networks.” Publication supported in part by an Institutional Grant (NA18OAR4170088) to the Texas Sea Grant College Program from the National Sea Grant Office, National Oceanic and Atmospheric Administration, US Department of Commerce. This
References (63)
- et al.
Examining branding co-creation in brand communities on social media: applying the paradigm of Stimulus-Organism-Response
Int. J. Inf. Manag.
(2018) - et al.
Social media for intelligent public information and warning in disasters: an interdisciplinary review
Int. J. Inf. Manag.
(2019) - et al.
Communicating on twitter during a disaster: an analysis of tweets during typhoon haiyan in the Philippines
Comput. Hum. Behav.
(2015) - et al.
Feasibility study of using crowdsourcing to identify critical affected areas for rapid damage assessment: hurricane Matthew case study
Int. J. Disaster Risk Reduct.
(2018) - et al.
Social network analysis: characteristics of online social networks after a disaster
Int. J. Inf. Manag.
(2018) Centrality in social networks conceptual clarification
Soc. Network.
(1978)Centrality in social networks conceptual clarification
Soc. Network.
(1978)- et al.
Models of core/periphery structures
Soc. Network.
(2000) - et al.
The role of social networks in natural resource governance: what relational patterns make a difference?
Global Environ. Change
(2009) - et al.
Institutional congruence for resilience management in interdependent infrastructure systems
Int. J. Disaster Risk Reduct.
(2020)
Including quality in Social network analysis to foster dialogue in urban resilience and adaptation policies
Environ. Sci. Pol.
Technology adoption and use in the aftermath of hurricane Katrina in new orleans
Am. Behav. Sci.
Backchannels on the front lines: emergent uses of social media in the 2007 Southern California Wildfires
The use of Facebook for information seeking, decision support, and self-organization following a significant disaster
Inf. Commun. Soc.
Government to Citizens (G2C) communication and use of social media in the post-disaster reconstruction phase
Environ. Hazards
Digitally enabled disaster response: the emergence of social media as boundary objects in a flooding disaster
Inf. Syst. J.
Social capital and community resilience
Am. Behav. Sci.
Ties that bond, ties that build: social capital and governments in post disaster recovery
Stud. Emerg. Order
The Cultural and Political Economy of Recovery: Social Learning in a Post-disaster Environment
Disaster risk reduction is not ‘everyone's business': evidence from three countries
Int. J. Disaster Risk Reduct.
Interorganizational resilience: networked collaborations in communities after superstorm Sandy
Soc. Netw. Anal. Disaster Response, Recovery, and Adapt.
Interorganizational coordination in dynamic context: networks in emergency response management
Connections
Twitter for crisis communication: lessons learned from Japan's tsunami disaster
Int. J. Web Based Communities
Social media use during Japa n's 2011 earthquake: how Twitter transforms the locus of crisis communication
Media Int. Aust.
Really social disaster: an examination of photo sharing on twitter during the #SCFlood
Vis. Commun. Q. Routledge
Network structure and community evolution on Twitter: Human behavior change in response to the 2011 Japanese earthquake and tsunami
Sci. Rep.
Exploring the emergence of influential users on social media during natural disasters
Int. J. Disaster Risk Reduct.
Robust detection of extreme events using twitter: worldwide earthquake monitoring
IEEE Trans. Multimed.
Integration of social media and unmanned aerial vehicles (UAVs) for rapid damage assessment in hurricane matthew
DUET: data-driven approach based on latent dirichlet allocation topic modeling
J. Comput. Civ. Eng.
Assessing Disaster Impacts on Highways Using Social Media: Case Study of Hurricane Harvey
Cited by (38)
Implications of 5G rollout on post-earthquake functionality of regional telecommunication infrastructure
2024, Journal of Infrastructure Intelligence and ResilienceAchieving fine-grained urban flood perception and spatio-temporal evolution analysis based on social media
2024, Sustainable Cities and SocietyExamining the consistency between geo-coordinates and content-mentioned locations in tweets for disaster situational awareness: A Hurricane Harvey study
2022, International Journal of Disaster Risk ReductionSocial media use in disaster recovery: A systematic literature review
2022, International Journal of Disaster Risk ReductionCitation Excerpt :Analysis of social media use during the response and recovery phases following Hurricane Harvey, revealed communication gaps between government organisations and citizens, which can be improved through regular updates, including information on infrastructure status, availability of shelters for citizens, and volunteering activities [75,76]. Though generally more active than government organisations, users from non-government organisations often participate through retweeting, mentioning, and commenting on disaster-related tweets of other organisations, particularly those of government organisations, instead of posting by themselves [75]. Liu et al. [77] analysed the Twitter networks and communication of 67 government and emergency management organisations to understand how they engaged with the public during and after Hurricane Harvey.
Text mining hurricane harvey tweet data: Lessons learned and policy recommendations
2022, International Journal of Disaster Risk ReductionCitation Excerpt :Increasingly, people are relying on cellular and internet-based technology to fulfill this need. For instance, disaster victims use social media to share their condition with friends and family [23,24]; and information about disaster magnitude, impacts, and needed assistance with relief organizations [16,21,25]. People also use social media to stay apprised of the extent of a disaster [26], to self-mobilize [23,27], and to seek emotional support and healing [28].
- 1
Akhil Anil Rajput and Qingchun Li have equal contribution to this paper.