Abstract
This study introduces a data visualization platform coupled to an unsupervised machine learning technique that has been developed using Google Earth Engine and OpenRailwayMap API to serve as a practical application for identifying potentially soft track segments. The goal is to provide maintenance-of-way engineers with a shortlist of locations that are highly likely to be soft, for a closer human inspection. The visualization platform is an interactive map that has been developed using Google Earth Engine by creating additional layers on Google Maps. OpenRailwayMap is also connected to this visualization platform as an independent layer to facilitate identifying rail structures. The results show that the cloud-based hybrid solutions (sensor measurements data, machine learning, and visualization) have the potential to fully automate the identification of track irregularities.
