Modeling an Information Network for Social Learning: A Case Study of Hurricane Ida Using Twitter Data
Dr. Yilan Xu
Associate Professor
The University of Illinois at Urbana-Champaign
In this paper, we use Twitter as a social sensor to help measure cross-region information
connectedness and construct an information network for social learning. Using Hurricane Ida as
a case study, we construct a weighted and directed climate disaster information network
described by an information network adjacency matrix. We further analyze the relative
importance of disaster-specific factors, demographics, and pre-existing social networks in
explaining the climate disaster information generation, diffusion, and network structure. We
also compare the disaster information network structure to several social network structures
including friendship, mobility, and migration. Our analytical framework can be generalized to
information diffusion on other topics, at different geographic scales, and using other social
sensors. The modeled disaster information network can contribute to the development of
effective disaster management strategies informed by real-time data. The information network
structure can be linked to behavioral responses to understand social learning in a human
network.