Getting Disaster Data Right: A Call for Real-time Research in Disaster Response
1 like80 views
1) The document describes a case study of various organizations that collaborated on disaster response efforts during the 2015 Nepal earthquake. This included using volunteers to analyze social media data and identify urgent needs. 2) It discusses challenges of conducting research during disasters and questions whether the Nepal model could be applied to future studies. This includes identifying mutually beneficial research topics and whether similar results could be obtained. 3) It proposes the Disaster Research Response repository to facilitate future research collaboration between practitioners and researchers during disasters.
Getting Disaster Data Right: A Call for Real-time Research in Disaster Response
- 1. Getting Disaster Data Right: A Call for Real-time Research in Disaster Response May 23, 2018 Steve Peterson, National Institutes of Health Chris Thompson, Humanity Road Cat Graham, Humanity Road ISCRAM 2018 The 15th International Conference on Information Systems for Crisis Response and Management
- 2. Agenda • Case Study • Correlate experiences to past research findings • Recommendations and proposed strategies for conducting future research during disasters
- 4. April 25, 2015 • Magnitude 7.8 Nepal earthquake – Killed ~9,000 people – Epicenter: Barpak, a village east of the Gorkha District Niranjan Shrestha/AP Images April 25, 2015 7.8 earthquake
- 5. Kathmandu Living Labs (KLL) • Initiated QuakeMap to serve as an information hub for public relief urgent needs – 4,000+ worldwide remote contributors plotted data • Sought crisis mapping assistance & urgent needs identification and coordination April 25, 2015 7.8 earthquake
- 6. Americares • Mobilized its medical teams to provide assistance & assess status of damage to health infrastructure • Sought assistance in identifying missing hospital addresses and contact information & assessing health infrastructure damage April 25, 2015 7.8 earthquake
- 7. Humanity Road • Overall disaster response activities involved multiple mission-specific projects spanning 49 days • Situation reports • Social media monitoring/amplification • QuakeMap platform administrator • Data mining • Information gathering • Requested surge support assistance to aid in providing updates to KLL & Americares April 25, 2015 7.8 earthquake
- 8. Community Emergency Response Team (CERT) • Reviewed data for urgent needs, medical information, & operating field hospital locations over a 48 hour period April 25, 2015 7.8 earthquake
- 9. About CERT • 500+ volunteers • Managed by Montgomery County Maryland Fire & Rescue Services • 6 week CERT Basic Course Training Topics – Disaster Preparedness – Fire Suppression Strategies – Medical Operations Part I & II – Light Search & Rescue Operations – Psychology – Course Review and Disaster Simulation
- 10. About CERT: Introduction of Technology Core Capability
- 11. Task Assignments • Keyword determination & search with social media analytics tool • Gather information on pre-identified hospitals and villages Task assignments changed midway through virtual activation • Seek urgent need requests and pertinent information on villages in three priority districts
- 12. Team Coordination and Communication • Volunteer management system email (20 volunteers responded) • Daily conference calls – Collective decisions based on situational awareness1 • Resource updates to Google document 1Van de Walle, B., Brugghemans, B., and Comes, T. (2016). Improving situation awareness in crisis response teams: An experimental analysis of enriched information and centralized coordination. International Journal of Human-Computer Studies, 95, pp. 66-79.
- 13. Data Collection and Analysis Model Team Lead Oversight Static Data (Medical-related keywords, priority districts, villages, and hospital addresses) Team 1 Dynamic Data (Emerging urgent needs and hospital statuses) Team 2 Consolidation of analyzed data Continuous leadership involvement supported test results of a 2011 study2 suggesting virtual team leaders, during crises, be more active to increase team members motivation3. 2Jenster, N.P. and Steiler, D. (2011). Turning Up the Volume in Inter-personal Leadership: Motivating and Building Cohesive Virtual Teams during Times of Economic Crisis. Advances in Global Leadership, pp. 267-297. 3Yu, X. and Khazanchi, D. (2017). Studying Virtual Teams during Organizational Crisis from a Sociomaterial Perspective. Proceedings of the 14th ISCRAM Conference, Albi, France. May 2017. pp. 1055-1064.
- 14. Information System and Human Data Analysis • Humanity Road provided access to a social media analytics information system called Scanigo – Improved ability to refine data into more manageable workloads, thus minimizing risk of information overload4 1.2 million unfiltered earthquake‐related tweets, retweets, shares, & posts Filtering/Categorization 213,000 medical‐related categorized tweets, retweets, shares, & posts Ranking 2,132 medical‐related ranked tweets, retweets, shares, & posts Human Data Analysis 41 urgent needs SM Analytics Toolkit SM Analytics Toolkit 4Gonzalez, J. J., Labaka, L., Hiltz, S. R., and Turoff, M. (2016). Insights from a Simulation Model of Disaster Response: Generalization and Action Points. 2016 49th Hawaii International Conference on System Sciences (HICSS).
- 15. Social media post Continue to next social media post No Unverified? Rumor/Fake news? No or Uncertain Official agency already respond? Yes No Clarifying facts available? Yes Yes Yes Location based on photo etc.? Yes Continue to next social media post Yes Continue to next social media post Outdated Relevant to ESF* assignment? Urgent? Yes Accurate & not outdated? Yes Yes (notify team lead) Team Collaborative Working Document No or Uncertain Useful & from credible source No Photo, video, audio? No or Uncertain Actionable IntelligenceVerified Report Team Lead Analysis Potential value to assignment? No Continue to next social media post No Continue to next social media post No Yes Human Data Analysis (SM Verification Steps) Social media post Not all inclusive Actionable IntelligenceVerified Report Team Lead Analysis Continue to next social media post No Unverified? Rumor/Fake news? No or Uncertain Official agency already respond? Yes No Clarifying facts available? Yes Yes Yes Location based on photo etc.? Yes Continue to next social media post Yes Continue to next social media post Outdated Relevant to ESF* assignment? *ESF: Emergency Support Function Urgent? Yes Accurate & not outdated? Yes Yes (notify team lead) Team Collaborative Working Document No or Uncertain Useful & from credible source No Photo, video, audio? No or Uncertain Potential value to assignment? No Continue to next social media post No Continue to next social media post No Yes
- 16. Verified Data Distribution Process
- 17. Results • 200+ volunteer hours • Health & medical keywords (413) • Medical-related urgent needs (41) • Hospital status updates (100) ~60% completed • Health & medical information on Village Development Committees (85) ~40% completed
- 18. Discussion
- 19. Discussion Questions • Could the Nepal virtual activation model be applied to researcher planning for future studies? How much outdated? How fast? How fast from time discovered to action taken? Stayed within lane? During data analysis period, known fake news/rumors may have been identified leading to analyst to quickly ignore investing further time. Did they? Did analyst make any “no” decisions based off simple conclusion w/o a degree of skepticism? Were they skeptical of their own skepticism?
- 20. Discussion Questions • Can practitioner-researcher consensus be achieved to identify mutually beneficial measurable topics for future research consideration during disasters? • Would conducting research within a time-sensitive accelerated environment, using the same methods of research conducted post- disaster, generate similar or different results? • Can the same prior research study conducted post-disaster also be applied during disasters in an effort to garner previously unknown, undocumented intelligence?
- 21. Discussion Questions • Can research methods and data collection strategies be applied to practitioners’ decision-making needs? • Can social media disaster research be performed during a disaster which generates results that convince skeptical practitioners of social media applicability and value? Available Research Resources: Volunteers
- 22. Disaster Research Response (DR2) Repository & Website Website: http://dr2.nlm.nih.gov Contact: dr2@nih.gov
- 23. Thank You Thank you! Steve Peterson, CEM Emergency Management Specialist National Institutes of Health steve.peterson@nih.gov Chris Thompson President Humanity Road chris@humanityroad.org Cat Graham Director of Operations Humanity Road cat@humanityroad.org