Big data analyses for real-time tracking of risks in the mineral raw material markets: implications for improved supply chain risk management

Abstract

Abstract Resilient supply chains have become a key issue for manufacturing companies to ensure a stable supply for their manufacturing processes and for governments to ensure the stable supply of essential goods to society. Building diversified supply chains and monitoring the performance of suppliers are major mitigation strategies to counteract disruptions at an early stage. Supply chain risk management and monitoring of supply chains using big data analytics are getting increasing attention. The growing number of data sources has huge implications on the reporting of incidents that may disrupt supply chains. The data sources may stem from a variety of internet sources like daily media reports provided on websites, social media or specialist media, or they may stem from specific databases. The sooner this information is disclosed to stakeholders and analysed the better the preventive strategies generally are. Timely information prolongs the reaction time and may help to reduce the severity of an incident. This paper highlights a science-based real-time tracking analysis of risks in the mineral raw material markets for the period 2019 to 2021 using big data analytics provided by a commercial system. A set of data for 12 selected mineral raw materials was provided by the authors and analysed using more than 100 risk indicators from 14 major risk categories as part of a commercial big data system. The extracted information can have imminent value to identify supply shortages, production halts or price peaks at an early stage. The main question was to find out whether such big data analytics are precise enough to detect potential, globally relevant, supply shortages in mineral raw material markets in due time. The results of this paper show that using big data analytics can be a very effective tool to extract relevant information about supply sources and to react timely in case of disruptions or social or environmental mismanagement on the supplier side. However, the nature of big data sources suggests that the parameters of the applied models need elaborate configuration. Each raw mineral market has its own peculiarities in terms of volume, mode of transport, market concentration or countries of origin. These factors influence the relevance of the reported incidents. Furthermore, some incidents have a spurious or only minor connection to the individual markets. For these reasons, we conclude that only supervised models reap the most benefits in the monitoring of mineral raw material markets.