Parallel Computing Approach for Preprocessing AIS Data in Transshipment Feature Extraction

AIS (Automatic Identification System) data is an important data source in the shipping and logistics industry to track and monitor ships in real-time. Extraction of transshipment features from AIS data is essential in understanding the pattern and characteristics of transshipment activity at a particular port or place. This study discusses an efficient AIS data preprocessing framework to prepare the transshipment feature extraction. The preprocessing steps include data collection, data cleansing, trajectory extraction, trajectory cleaning, and overlapping trajectory extraction. One of the crucial approaches used in this study is the application of parallel computing. Parallel computing allows us to execute independent data processing tasks at the same time simultaneously. In AIS data preprocessing, the parallel approach allows us to clean, populate, and process data from many vessels in less time. To test the applicability, we created a final comparison for different types of ships based on its AIS data for Indonesia from November 2022 to April 2023. Using parallel computing to process AIS data makes data processing more efficient, especially for large and complex AIS data. This efficiency accelerates data preparation before transshipment feature extraction. The results show that the framework can accelerate runtime and improve trajectory quality.