AIS Data Mining
A Pipeline for Pre-Processing, Cleansing, and Visualizing AIS data with Ports Network Generation
During my M.Sc. at the University of Pisa, I had the opportunity to work as a visiting student at the Institute for Big Data Analytics at Dalhousie University. I focused on AIS data for cruise ships. After developing a pipeline for preprocessing, cleansing, and generating a network of ports in Python for the ship data, I designed a dashboard to visualize the Automatic Identification System (AIS) trajectories of vessels. Given the high volume of points characterizing these trajectories, efficient visualization was crucial. To achieve this, I applied the Douglas-Peucker line simplification algorithm to simplify the polylines of trajectories. Subsequently, I plotted trajectories at various levels of detail. When zoomed out, a coarse representation of trajectories is displayed, while zooming in reveals finer details. I incorporated this functionality by extending the Polyline class of Leaflet.js. Specifically, when in a zoomed-out view and the number of points falls below a predefined threshold, the trajectory points are displayed. Otherwise, only the polylines are shown, and the points are discarded.
