Application of graph methods for leveraging open source data during disaster response

Abstract

Every disaster is unique, yet all response efforts face common challenges regarding the collection and processing of disparate data sources, and dissemination of accurate and timely information both to and from disaster areas. Recent advances in graph theory and data fusion can be leveraged to address many of these challenges. This paper describes a graph-based approach for fusing and representing multisource information of an affected area to aid emergency responders in the wake of a disaster. An end-to-end prototype system was developed, consisting of five main parts: data ingestion, social graph construction, graph enrichment, inference, and user situation awareness. Data from open sources, including social media, are ingested and fused to represent people and places as a coherent social graph. This graph can be made available to emergency managers and incident commanders to increase situation awareness of an affected population, or used as inputs to other algorithms to aid in prioritizing a response. Key challenges of using open source data for disaster response are discussed.