A network science approach to open source data fusion and analytics for disaster response

Network science is often used to understand underlying phenomena that are reflected through data. In real-world applications, this understanding supports decision makers attempting to solve complex problems. Practitioners designing such systems must overcome difficulties due to the practical limitations of the data and the fidelity of a network abstraction. This paper explores the design of a network science solution for the disaster relief domain with the goal of increasing the efficiency of disaster response efforts. Various real-world network science challenges are discussed relating to entity disambiguation and relationship estimation as well as general data science challenges such as limited access to representative data and learning inference models in this environment. A novel graph-based information management system was designed and prototyped to access and aggregate data from multiple sources. The system consists of five main parts: data ingestion, graph construction, inference, situational awareness, and evaluation. Data from open sources, such as social media, are ingested and fused to represent people, places, and social media users as a coherent social graph. This graph can be displayed to first responders to increase situational awareness or used as inputs to algorithms for graph analytics that support response efforts. Due to the lack of historical data from disaster events, an agent-based simulation was developed to create representative social graphs.