System Safety Data Network: Architecture and Blueprint

Abstract

With increasing complexity of safety analysis in sociotechnical systems, there is a need for a mechanism to accurately capture complex information and present it in an easily accessible and understandable form. While there are plenty of accident databases that have been created over the years for specific purposes, a tool that provides a holistic view of all the safety-related aspects of an accident customized specifically per user and industry is largely absent. This paper discusses the conceptual model of the system safety database (SSD), a tool that will offer tailored solutions to multiple classes of users and that will generate reports synthesizing lessons learned from a variety of disparate contexts, providing succinct and actionable information for decision support. The paper also proposes the concept and architecture of a System Safety Data Network (SSDN) that encapsulates a network of safety databases, thereby addressing some of the challenges of a stand-alone safety database. The data network will enable working with structured and unstructured data by integrating multiple relational and NoSQL databases. A full-fledged implementation of the SSDN will enable improved collaboration across industries and corporations. The System Safety Data Network will facilitate analysis across disciplines and contexts, allowing researchers and practitioners to use integrated mixed-methods approaches to conduct investigations, analyses, research, and development activities across multiple levels of a system. The paper also discusses the steps involved in the implementation of such a data network and the challenges involved. In addition, the current work in data categorization and interpretability of incident data is discussed. When completed, the System Safety Data Network will provide stakeholders at all levels, from individual operators to policymakers, with the tools and perspectives needed to improve the safety of complex sociotechnical systems.