Matt Himes — Project Report
Area of Investigation: Disaster relief coordination. As we have seen yet again with the debacle in Puerto Rico, organizing and coordinating disaster relief efforts is a massive and incredibly difficult undertaking. I want to design a cognitive system that can help disaster relief organizations (both government and NGO) get humanitarian aid to more people faster and more efficiently than humanly possible.
Key problem: Coordination and lack of organizational capacity. My secondary research has led me to alter my hypothesis slightly, to focus more on using A.I. to address logistical coordination issues, so that the many different humanitarian aid organizations that respond to disasters avoid duplicating efforts, competing over resources, and overall function more effectively.
Target user: Denise the Disaster Relief Coordinator.
Justification for cognitive: I see cognitive having the ability to combine a variety of data sets, including weather reports, satellite imaging, road map data, crowdsourced relief requests, reports from first responders on the ground, and information from various relief organizations, to paint a more complete picture of the emergency in the form of a constantly evolving crisis map. This will allow the various different organizations responding to the crisis to pool information and resources, work off of shared information, track where aid is needed and what has already been send, and ultimately make better use of humanitarian aid agencies through better collaboration. Machine learning could also be used to ID tweets about a crisis and crowdsource relief requests directly from the people affected.
Research Proposal
Background: There were over 250 natural disasters between 2005 and 2016, including floods, fires, earthquakes, droughts, hurricanes, landslides, and floods (source). We have also seen the devastation caused by hurricanes Harvey, Irma, and Maria over just the last few months. As we have seen from the recent challenges in Puerto Rico, the logistical efforts required to organize relief and humanitarian aid campaigns for events this large are absolutely massive and overwhelming. One of the largest challenges seems to be the fact that there are many different organizations, both government and NGO, that each respond to a crisis in their own way. This leads to the various aid agencies stepping on eachother’s toes rather than collaborating to optimize recovery efforts. This seems like a great opportunity to utilize machine learning because where there is too much data for human agents to comprehend, intelligent systems can help synthesize huge amounts of data into digestible forms better than humanly possible. By combining artificial intelligence and human intelligence, relief efforts can be optimized and more lives can be saved.
What has been done: While I found a variety of articles and papers about how A.I could potentially be used to assist in disaster recovery efforts, I could only find one that is currently on the market. It is called 1Concern, and it seems to try to address many of the same issues I have identified here, as well as many many others. From what I can tell it is a huge platform with a broad range of capabilities, but not much detail available without requesting a demo. There have also been conferences on this issue, including one called ISCRAM (Intelligent Systems for Crisis and Disaster Management) that dates back to at least 2008 (source).
Research Objective: To understand how to help different humanitarian aid organizations can better coordinate to optimize disaster recovery efforts.
Research Participants: Humanitarian aid organizations, specifically people involved with coordination of relief efforts.
Methodology: In addition to secondary research, I will be reaching out to different humanitarian aid organizations (FEMA, Oxfam, RedCross, WorldVision, SavetheChildren, etc.) to try and conduct short phone interviews. The interviews will focus on identifying major logistical and organizational challenges to recovery efforts, and how thee efforts could be better optimized.
Secondary Research:
Sources Cited:
1. Challenges and obstacles in sharing and coordinating information during multi-agency disaster response: Propositions from field exercises. Bharosa, N., Lee, J. & Janssen, M. Inf Syst Front (2010) 12: 49. https://link.springer.com/article/10.1007%2Fs10796-009-9174-z
2. “Providing Disaster Relief An Ongoing Challenge.” Talk of the Nation. National Public Radio. September 21, 2010. Radio. http://www.npr.org/templates/story/story.php?storyId=130021483
3. “Relief Challenges.” Inside Disaster Haiti, insidedisaster.com/haiti/response/relief-challenges. Accessed 28 Sept. 2017. http://insidedisaster.com/haiti/response/relief-challenges
4. Gao H., Wang X., Barbier G., Liu H. (2011) Promoting Coordination for Disaster Relief — From Crowdsourcing to Coordination. In: Salerno J., Yang S.J., Nau D., Chai SK. (eds) Social Computing, Behavioral-Cultural Modeling and Prediction. SBP 2011. Lecture Notes in Computer Science, vol 6589. Springer, Berlin, Heidelberg. https://link.springer.com/chapter/10.1007/978-3-642-19656-0_29#citeas
5. “The Benefits & Challenges of Using Artificial Intelligence for Emergency Management.” Using Artificial Intelligence for Emergency Management | EKU Online, safetymanagement.eku.edu/resources/infographics/the-benefits-challenges-of-using-artificial-intelligence-for-emergency-management/. Accessed 28 Sept. 2017. http://safetymanagement.eku.edu/resources/infographics/the-benefits-challenges-of-using-artificial-intelligence-for-emergency-management/
6. “Groundbreaking approach to disaster relief.” WHO, World Health Organization, www.who.int/bulletin/volumes/86/9/08-010908/en/. Accessed 28 Sept. 2017. http://www.who.int/bulletin/volumes/86/9/08-010908/en/
7. Brandom, Arielle Duhaime-Ross and Russell. “Why are we still coordinating disaster relief over radios?” The Verge, The Verge, 14 May 2015, www.theverge.com/2015/5/14/8607371/disaster-relief-radio-ems-fire-police. Accessed 28 Sept. 2017. https://www.theverge.com/2015/5/14/8607371/disaster-relief-radio-ems-fire-police
8. H. Gao, G. Barbier and R. Goolsby, “Harnessing the Crowdsourcing Power of Social Media for Disaster Relief,” in IEEE Intelligent Systems, vol. 26, no. 3, pp. 10–14, May-June 2011. http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5898447&isnumber=5898441
9. Written by Demian Repucci, et al. “OI Engine, an innovation management software built on design thinking.” OpenIDEO — What global challenge do you think innovation leaders should work to solve right now? — How Can Disaster Relief be Coordinated to Make it Most Efficient And Effective?, 15 Dec. 2010, challenges.openideo.com/challenge/what-is-the-global-challenge-that-most-concerns-you-right-now-and-that-global-innovation-leaders-could-begin-to-solve/applause/how-can-disaster-relief-be-coordinated-to-make-it-most-efficient-and-effective. Accessed 28 Sept. 2017. https://challenges.openideo.com/challenge/what-is-the-global-challenge-that-most-concerns-you-right-now-and-that-global-innovation-leaders-could-begin-to-solve/applause/how-can-disaster-relief-be-coordinated-to-make-it-most-efficient-and-effective
Summary of Secondary Research: I have revised my hypothesis slightly to focus more on logistical problems and the lack of collaboration between aid agencies. Previously I had identified the problem as the fact that aid takes longer than it needs to to reach victims due to poor weather and road conditions, etc. After conducting my secondary research I realized that while weather and road conditions certainly play a part, the larger issue to address is the lack of communication and collaboration between various aid agencies resulting in competition for resources, duplication of efforts, and strain on organizational capacity.
Old Elevator Pitch: Denise the disaster relief coordinator can distribute aid faster and more efficiently by utilizing weather and road condition data to predict where resources need to be and when.
New Elevator Pitch: Denise the disaster relief coordinator can make better use of humanitarian aid resources by pooling information with other aid agencies and collaborating to optimize recovery efforts.
Primary Research
I have reached out to about a dozen different organizations at different levels (state, local, national), to try and get in touch with appropriate people to interview. In the meantime, I have written a first draft of my interview questions, and I would appreciate any feedback.
Interview questions: .
1. What is your title and job description?
2. How many disaster relief efforts have you been a part of?
3. Can you walk me through a typical day as a disaster relief coordinator?
4. What type of data is most important during a relief effort, and how is that data delivered to the people making decisions?
5. Is there any data that is not typically available that would help you do your job better?
6. How do you identify where where help is needed? What methods are used to communicate with victims?
7. How many other aid organizations do you typically interact/collaborate with during a relief effort?
8. How effective do you think that collaboration typically is? Would more collaboration between organizations equal a better/more efficient recovery effort?
8. Have you experienced any instances of organizations competing for resources, duplicating efforts, or generally stepping on each others toes?
9. What type of software is used to assist in relief coordination? Is there a common system or is everyone using their own software?
10. Are there any other changes you would recommend to try and make disaster relief more effective and efficient?
