How effective is social media for waterlogging disaster detection by twitter users?

The prompt and efficient delivery of data pertaining to the havoc wrought by calamities is crucial for the expeditious process of decision-making. The utilization of social media platforms has facilitated the distribution of crucial emergency information, which has been extensively examined as an indispensable instrument for obtaining and evaluating information pertaining to devastation. In this study, a framework is proposed for the swift categorization of damage and the monitoring of restoration efforts in the context of rainstorm and urban waterlogging. The framework leverages the potential of big data derived from social media, particularly Twitter.  Karachi cosmopolitan city witnessed a real-life instance of urban waterlogging between the months of June and October in the year 2022, the proposed framework is effectively demonstrated through the use of an illustrative case study. This research utilizes a novel approach that utilizes the integration of social media data, land use data, and other relevant data sources to examine the spatial and temporal patterns of public responses to urban rainstorm and waterlogging in a specified urban region from June 14 to October 2022. Spatially discernible clusters of elevated public concern with respect to urban waterlogging were observed within regions characterized by relatively low elevation, a significant proportion of water bodies, and a marked surge in urban development in recent years. Temporal fluctuations reveal that public reactions typically culminate during the rainstorm phase as opposed to preceding it. Notably, alterations in behavior and emotional states are most conspicuous during the morning rush hour within a 24-hour period, in connection with the rainstorm. The revelation of discrepancies in the response among individuals from different backgrounds and the impact of urban waterlogging on various aspects of individuals' daily lives have become evident. The framework proposed in this study offers a viable approach for acquiring on-site damage data prior to carrying out thorough assessments in regions impacted by calamities. This methodology facilitates the formulation of real-time management tactics for local flood prevention and the advancement of sustainable growth in areas susceptible to damaged areas caused by rainstorms.