The number of people vulnerable and affected by disasters has been increasing since the 1970s. Events like the 2010 earthquakes in Haiti and Pakistan, the 2011 Fukushima Daiichi nuclear disaster, and the 2013 typhoon Haiyan in the Philippines caused huge losses in terms of lives and economic damages. For the successful delivery of relief items during disaster response, humanitarian logistics makes use of a wide variety of infrastructure and resources, like the road network or vehicles. In the immediate aftermath of a disaster, much needed up-to-date and sufficiently accurate information about available infrastructure and resources is lacking most. What has been termed “big crisis data” – including data from social media platforms, such as Twitter and Facebook – has come into the focus of research as source of information about the unfolding situation in disaster contexts. However, research on the targeted extraction of information that can be fed back into operational processes is still in its infancy. A fundamental challenge lies in the combination of human and machine intelligence to quickly identify task-related information while avoiding information overload.

The main objective of this thesis is to support the labor-intensive content moderation by facilitating the identification of task-related information with SML for decision-making on humanitarian logistics, focusing on the information about infrastructure and resources collected from mobile devices and online social media. The foundation for this revision is twofold. On the one hand, the existing content moderation workflow involves human annotators and information categories to semantically enrich messages (see Link et al. 2013). A semi-automated process with SML for the moderation of crowdsourced messages from both mobile devices and online social media should be designed as an extension to the existing content moderation workflow. On the other hand, a newly developed method for keyword-based search of task-related information in social media data (Link et al., 2015) managed to link crisis-related messages to the domain-specific tasks of humanitarian logistics via information categories about infrastructure and resources, which provides a foundation for the development of enhanced SML-based classifiers that incorporate the domain-specific knowledge (in the form of existing information categories and previously generated keywords) for informative message filtering and auto-classification of crisis-related messages into the task-related information categories for decision making on humanitarian logistics. Social media data for testing can be downloaded from CrisisLex.org.

Recommended reading:

• 
  Horita, Flávio E. A.; Link, Daniel; Porto de Albuquerque, João; Hellingrath, Bernd (2014): A Framework for the Integration of Volunteered Geographic Information into Humanitarian Logistics. In AIS (Ed.): AMCIS 2014 Proceedings. Savannah, GA.
• 
  IFRC (2013): World Disasters Report 2013. Focus on technology and the future of humanitarian action. Edited by Patrick Vinck. International Federation of Red Cross and Red Crescent Societies. Geneva.
• 
  Imran, Muhammad; Castillo, Carlos; Diaz, Fernando; Vieweg, Sarah (2014): Processing Social Media Messages in Mass Emergency: A Survey. In ArXiv e-prints (1407.7071). Available online at http://arxiv.org/abs/1407.7071, checked on 2/13/2015.
• 
  Link, Daniel; Hellingrath, Bernd; Groeve, Tom de (2013): Twitter Integration and Content Moderation in GDACSmobile. In Tina Comes, F. Friedrich, S. Fortier, J. Geldermann, T. Müller (Eds.): Proceedings of the 10th International ISCRAM Conference. Baden-Baden, pp. 67–71.
• 
  Link, Daniel; Horita, Flávio E. A.; Albuquerque, João Porto de; Hellingrath, Bernd; Ghasemivandhonaryar, Shabdiz (2015): A Method for Extracting Task-related Information from Social Media based on Structured Domain Knowledge. In : Proceedings of the 2015 Americas Conference on Information Systems. AMCIS 2015. Fajardo, PR, 13.-15.08.2015. Association for Information Systems (AIS).
• 
  Meier, Patrick (2015): Digital humanitarians. How big data is changing the face of humanitarian response.