Big data and artificial intelligence (AI) are crucial components of data-driven decision-making in most industries (Liang and Liu 2018). The maritime industry is one of the oldest and traditional industries to still rely more on intuition than on data, due to the vast size of network and planning problems (Brouer, Karsten, and Pisinger 2016). Big data and AI have received considerable attention in recent years, through a number of publications, and some scholars have portrayed the concept of ‘big data’ as hype (d’Amore, Baggio, and Valdani 2015). The term big data is typically used to denote large amounts of data. With the recent burst of data volume, researchers have been continuously scrutinising novel techniques for analysing big data (Franks 2012). A branch of these techniques is now integrated into the concept of ‘AI’.
AI research initially aimed to mimic human decision-making by utilising a large volume of data using machines. Nowadays, AI is capable of doing things that were impossible a decade ago. For example, sophisticated AI systems introduce autonomous ships, which can operate independently without human interaction, and the error rate is lower than that of human-operated ships. AI is gradually transforming the traditional operational process of the maritime industry. Consequently, the amount of research on the application of big data and AI has increased significantly since 2012 (Liang and Liu 2018). Following this trend, data-centric innovative technologies and new business models are being developed (Munim 2019). This transformation is reshaping the maritime industry, providing new opportunities to improve productivity, efficiency and sustainability (Heilig, Lalla-Ruiz, and Voß 2017).
Studies on the synthesis of big data application in maritime are rare, which has created a gap in the academic literature due to the importance of big data and AI in maritime operations (Yang et al. 2019; Mirović, Miličević, and Obradović 2018). Big data- and AI-enhanced maritime operations can contribute to the economic and environmental aspects of the maritime business (Sanchez-Gonzalez et al. 2019). Maritime trade accounts for approximately 80% of world trade (UNCTAD 2018) and the industry faces many challenges due to its vastness (Brouer, Karsten, and Pisinger 2016) as well as continuously evolving regulatory requirements (Lee, Kwon, and Ruan 2019). Big data and AI offer viable solutions to some of these challenges. For example, data about ship performance and navigation systems can help shipping firms monitor vessels’ performance and take necessary steps to improve the operational efficiency of the vessels (Mirović, Miličević, and Ines 2018). The industry generates large amounts of data that, if appropriately utilised in decision-making, can improve maritime safety, reduce environmental impacts and minimise cost.
To the best of our knowledge, in the maritime context there have been two review studies on big data (Yang et al. 2019; Mirović, Miličević, and Ines 2018) and two on digitalisation (Sanchez-Gonzalez et al. 2019; Fruth and Teuteberg 2017). The present study is more comprehensive than previous studies in terms of quality and spread of included studies that use big data and AI in the maritime context. For instance, Yang et al. (2019) reviewed studies that use only automatic identification systems (AIS) data. Unlike the present study, Mirović, Miličević, and Obradović (2018) did not follow a systematic approach to literature selection, which can lead to biased findings. Fruth and Teuteberg (2017) and Sanchez-Gonzalez et al. (2019) explicitly focused on digitalisation, although both used big data in their keyword search. While Fruth and Teuteberg (2017) did not include AI aspects in the maritime domain, Sanchez-Gonzalez et al. (2019) literature search process is rather abstract and may not be reproducible.
Unlike previous review studies, the literature search process in the present study was robust, transparent and reproducible. We review published studies that deal with big data and AI applications within the maritime context to map the conceptual structure of the field and identify future research avenues. Hence, we address four research objectives. The first is to find the existence of the big data and AI research in maritime as a standalone research domain. The second is to identify the key journals, articles, institutions and authors within this research domain and find the collaborative network of universities and authors. The third is to map the conceptual structure of big data and AI research in maritime by identifying and exploring underlying research clusters. The final objective is to extract and present the avenues for future research.
The findings of this study have several academic and industry implications. For the scholars and practitioners interested in big data and AI research in maritime, it provides a comprehensive overview of the research domain that introduce readers with the key studies, authors, universities, concepts and methods. Maritime firms and regulatory authorities can use the identified concepts and methods to enhance coordination among major players, optimise resource use, and improve environmental performance and navigational safety.
Comments