The objective of this analysis is to find hidden patterns in vessel behavior. Illegal, Unreported, and Unregulated (IUU) fishing often involves vessels turning off automatic identification system (AIS) trackers or moving in erratic patterns to avoid detection. By processing geospatial data, using K-Means clustering to segment vessels into groups based on their movement patterns, we'll be able to identify anomalous clusters vs normal fishing behavior. The analysis includes: unsupervised learning (clustering), geospatial data handling, visualization, and pattern recognition. The two datasets used are from the Global Fishing Watch organization, specifically their Anonymized AIS Training dataset, and Welch et al. (2022).Â
The algorithm successfully isolated an anomalous cluster exhibiting physically impossible average speeds (44+ knots) and erratic movements. This subset strongly indicates intentional AIS/GPS manipulation to mask true locations, separating them from normal transiting or active fishing fleets.
By visualizing these anomalies, authorities can shift from manually monitoring oceans to investigating targeted high-risk zones. Future work involves cross-referencing these areas with Marine Protected Areas.