Spatio-temporal filtering techniques for the detection of disaster-related communication.
Individuals predominantly exchange information with one another through informal, interpersonal channels. During disasters and other disrupted settings, information spread through informal channels regularly outpaces official information provided by public officials and the press. Social scientists...
| Published in: | Social Science Research Vol. 59; pp. 137 - 155 |
|---|---|
| Main Authors: | , , , |
| Format: | Article |
| Published: |
Academic Press Inc.
Sep2016
|
| Subjects: | |
| Online Access: | View this record in EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=117269812&site=ehost-live header: @attributes: shortDbName: ssf uiTerm: 117269812 longDbName: Social Sciences Full Text (H.W. Wilson) uiTag: AN controlInfo: bkinfo: jinfo: jid: 0049089X SSS jtl: Social Science Research issn: 0049089X maglogo: N pubinfo: dt: Sep2016 vid: 59 pid: 735 pub: Academic Press Inc. artinfo: ui: 117269812 10.1016/j.ssresearch.2016.04.023 ppf: 137 ppct: 18 formats: tig: atl: Spatio-temporal filtering techniques for the detection of disaster-related communication. aug: au: Fitzhugh, Sean M. Ben Gibson, C. Spiro, Emma S. Butts, Carter T. affil: Department of Sociology, University of California, 3151 Social Science Plaza A, Irvine, CA 92697, USA Information School, Mary Gates Hall, 370, University of Washington, Seattle, WA 98195, USA Institute for Mathematical Behavioral Sciences, Department of Statistics, Bren Hall 2019, University of California, Irvine, CA 92697, USA Department of Electrical Engineering and Computer Sciences, 2200 Engineering Hall, University of California, Irvine, CA 92697, USA su: Social scientists Information filtering Emergency communication systems Spatiotemporal processes Spatial data structures sug: subj: Social scientists Information filtering Emergency communication systems Spatiotemporal processes Spatial data structures keyword: Big data Communication Disasters Event detection Geography Rumoring Big data Communication Disasters Event detection Geography Rumoring ab: Individuals predominantly exchange information with one another through informal, interpersonal channels. During disasters and other disrupted settings, information spread through informal channels regularly outpaces official information provided by public officials and the press. Social scientists have long examined this kind of informal communication in the rumoring literature, but studying rumoring in disrupted settings has posed numerous methodological challenges. Measuring features of informal communication–timing, content, location–with any degree of precision has historically been extremely challenging in small studies and infeasible at large scales. We address this challenge by using online, informal communication from a popular microblogging website and for which we have precise spatial and temporal metadata. While the online environment provides a new means for observing rumoring, the abundance of data poses challenges for parsing hazard-related rumoring from countless other topics in numerous streams of communication. Rumoring about disaster events is typically temporally and spatially constrained to places where that event is salient. Accordingly, we use spatio and temporal subsampling to increase the resolution of our detection techniques. By filtering out data from known sources of error (per rumor theories), we greatly enhance the signal of disaster-related rumoring activity. We use these spatio-temporal filtering techniques to detect rumoring during a variety of disaster events, from high-casualty events in major population centers to minimally destructive events in remote areas. We consistently find three phases of response: anticipatory excitation where warnings and alerts are issued ahead of an event, primary excitation in and around the impacted area, and secondary excitation which frequently brings a convergence of attention from distant locales onto locations impacted by the event. Our results demonstrate the promise of spatio-temporal filtering techniques for “tuning” measurement of hazard-related rumoring to enable observation of rumoring at scales that have long been infeasible. pubtype: Academic Journal doctype: Article src: R language: English refInfo: copyright: @attributes: flag: N holdings: @attributes: islocal: N |
|---|