This work studies how the tasks performed by organizations responding to a large-scale disaster affects the structures and interactions within an emergent multiorganizational network (EMON) of collaborations. We formulate a large language model-based methodology for systematic extraction of organizational task structures from long-form situational reports detailing the activities of organizations responding to the 2005 Hurricane Katrina disaster. Tasks are related to a broader disaster response task ontology, then used in exploratory multivariate analysis augmented by nonparametric hypothesis tests to understand how the tasks performed by each organization relate to their collaborations and positions within the broader EMON structure. • Membership in the giant component of collaborations is associated with higher average levels of task performance. • High degree nodes, on average, perform larger numbers of tasks but some isolated nodes are very active in task performance. • Node position in the collaboration network and the types of tasks performed are associated. • Work on related tasks is positively associated with collaborative relationships.
Mai et al. (Sat,) studied this question.