This presentation was prepared for the SPATRA webinar on the rail workstream, held on 21 May 2025. The slides explain the SPATRA approach for using satellite-based Earth Observation (EO) data and machine learning to support rail infrastructure safety. Two main topics are covered: Using Satellite-Based EO Data and Machine Learning to Predict Rail Track Temperatures, presented by Alina Klapper (OHB Digital Services GmbH) Machine-Learning-Based Prediction of Rail Track Buckling Using Temperature Data Derived from Satellite Earth Observation, presented by Milan Banić (Faculty of Mechanical Engineering, University of Niš) The presentation describes data sources (Sentinel-3, MODIS, Copernicus), downscaling methods, and predictive models used to estimate rail track temperatures and identify buckling risk areas. These results help improve rail maintenance planning and reduce safety risks caused by extreme temperatures. Webinar recordings are available on SPATRA YouTube Channel: https://youtu.be/njFHX4w-nA8 More information on Rail use case is available on the SPATRA project website: https://spatra-project.eu/use-case-2-rail/