Abstract

The proliferation of sensor-generated data offers new possibilities for improving the way official statistics are produced, particularly in the maritime sector where timely, accurate, and granular information is increasingly important for evidence-based planning and policy evaluation. Maritime statistics have traditionally depended on administrative records and surveys, which often face challenges related to reporting delays, limited coverage, and respondent burden. Against this background, this study proposes a conceptual and methodological framework for integrating Automatic Identification System (AIS) data as an alternative and complementary source for official maritime statistics.
The proposed framework describes how large volumes of vessel movement data can be transformed into meaningful maritime indicators through the combined use of big data analytics and artificial intelligence, while embedding explicit quality assurance (QA) steps for AIS as the input data. In line with a big-data QA approach, the framework applies AIS data sourced from the United Nations Global Platform (UNGP) and implements QA prior to downstream processing and analysis. QA is structured around three pillars including (1) developing an input-data quality matrix that flags quality issues rather than simply removing records, (2) preprocessing (data cleaning and selection of relevant records) to produce an analysis-ready dataset, and (3) business process understanding to validate derived outputs against real-world operational conditions.
Rather than focusing on a single port or operational setting, the framework is designed to be flexible and scalable across geographic contexts and institutional environments. By exploiting the spatio-temporal characteristics of AIS data, the approach enables a more detailed understanding of maritime activity, port dynamics, and related economic and environmental dimensions that are often difficult to capture using conventional methods. As an illustration of statistical deliverables, the QA and processing pipeline is oriented toward producing AIS-based indicators such as port call statistics, cargo estimation, and vessel ownership identification.
In addition to methodological considerations, the paper discusses broader implications of adopting innovative data sources for official statistical systems, including data governance, institutional capacity, interoperability, and alignment with established quality standards. The quality matrix is developed by drawing on the Statistical Quality Assurance Framework (QAF) dimensions used by BPS-Statistics Indonesia, covering aspects such as relevance, accuracy, timeliness/punctuality, coherence, comparability, and trustworthiness, to support transparent and credible AIS-based statistics.