Data Sparseness in Container Terminals
3 Proven Methods for Ports to Overcome Data Gaps
In modern port operations, reliable and complete data is essential for efficiency, sustainability, and smart decision-making. Yet, many container terminals face data sparseness, which refers to gaps or inconsistencies in operational data caused by system limitations, manual overrides, or fragmented information between stakeholders.However, the availability of large amount of high quality terminal and equipment data is not a given in a complex operation. For example, tracking containers movements in a port is challenging due to system limitations, manual overrides, and data silos between stakeholders.
This blog gives an overview of three ways of generating data in case of data sparseness at container terminals and how to cope with it:
1: Interpolation for Filling Missing Data Points in Port Operations
When there is sparse data or only a few data points, interpolation can be used to estimate the intermediate data point. The relationship between different data points can be determined, for example, a linear relationship between x and y. This method works well for linear system where only the amount of data is an issue. In case of complex systems, such as container terminals, the relationship between different data points is difficult to estimate. For example, moving one container may require additional shuffle moves, depending on vessel schedules and yard designs, making it a non-linear, complex system. In the case of poor data quality (such as noise due to human errors or bias due to overrepresentation of the same values in the data), the estimation of the data will also be poor. Interpolation is a useful method for coping with data sparseness at container terminals in case of missing data points for linear systems.
2: Creating Synthetic Data Using AI for Container Terminals
AI algorithms can be used to create synthetic data based on a small set of the original data. Based on the characteristics and structure of this small set, AI learns and can develop a new set of data to compliment with the original data. This is useful to gather more data on non-linear system, such as container operations, and can be used as input for analysis. However, if the original terminal data is of poor quality (noise or bias), the AI algorithm will also create poor data as it learns from the original data. Furthermore, AI algorithms are often a blackbox, meaning it is challenging to validate the reliability of the synthetic data. Hence, creating synthetic data using AI is useful when data quality is guaranteed.
3: Using Simulation Modelling to Overcome Data Gaps in Logistics
Simulation modelling combines data with causal relationships, and the behaviour of complex systems using experiences, scientific theories, and universal laws. By combing data and cause-and-effect relationships, it is possible to simulate a complex system and obtain data on elements on which no data is available currently.
However, a simulation model needs data as input – it being limited or containing noise or bias. To understand the complex system with the uncertainty of missing / poor data, it is necessary to consider different models of the reality – the many model principle. By using different parameters and structural components in a computer-driven way, many different possible simulation models of the future can be generated automatically. This gives us insight into multiple plausible models of how the system could work, and we can test different “what-if” scenarios on this set of models. Hence, the bottlenecks across a set of plausible futures can be identified or the robustness of actions can evaluated.
This method does not rely solely on the (missing / poor) data but combines it with causal and factual relationships. In addition, it considers a set of possible models of reality rather than a single one, leading to robustness. Also, simulation models are often insightful, making the behaviour and trends – and thereby the data – easier to validate and verify with experts.
If your port or container terminal is facing data sparseness, Portwise can help you identify the best approach, ranging from interpolation to advanced simulation modelling, to make robust and futureproof decisions. Let’s explore how simulation and strategic consulting can help you invest with confidence and play a vital role in supporting progress towards global circularity and net-zero targets. Book a Free Consultation
About the author:
Dr. Isabelle van Schilt works as Project Manager at Portwise. Her expertise focuses on dealing with (deep) uncertainty, data sparseness, and robust decision-making in the ports, logistics, and supply chain industry. She holds a PhD degree in supply chain simulations from Delft University of Technology.