Challenges, opportunities and paradigm of applying big data to production safety management: From a theoretical perspective

Big data has caused the scientific community to re-examine the scientific research methodologies and has triggered a revolution in scientific thinking. As a branch of scientific research, production safety management is also exploring methods to take advantage of big data. This research aims to provide a theoretical basis for promoting the application of big data in production safety management. First, four different types of production safety management paradigms were identified, namely small-data-based, static-oriented, interpretation-based and causal-oriented paradigm, and the challenges to these paradigms in the presence of big data were introduced. Second, the opportunities of employing big data in production safety management were identified from four aspects, including better predict the future production safety phenomena, promote production safety management highlight relevance, achieve the balance between deductive and inductive approaches and promote the interdisciplinary development of production safety management. Third, the paradigm shifting trend of production safety management was concluded, and the discipline foundation of the new paradigm was considered as the integration of data science, production management and safety science. Fourth, a new big-data-driven production safety management paradigm was developed, which consists of the logical line of production safety management, the macro-meso-micro data spectrum, the key big data analytics, and the four-dimensional morphology. At last, the strengths (e.g., supporting better-informed safety description, safety inquisition, safety prediction) and future research direction (e.g., theory research focuses on safety-related data mining/capturing/cleansing) of the new paradigm were discussed. The research results not only can provide theoretical and practical basis for big-data-driven production safety management, but also can offer advice to managerial consideration and scholarly investigation. (C) 2019 Elsevier Ltd. All rights reserved.