This thesis presents a data-driven initiative developed during an internship within the Global Engineering Services (GES) organization at Amazon, aimed at improving project timeline accuracy and resource allocation efficiency across a geographically distributed portfolio of infrastructure projects. The work addresses a fundamental challenge in large-scale program management: the absence of a standardized and empirically validated framework for project planning and timeline estimation. To tackle this problem, the research follows a structured and progressive methodology, articulated across three interconnected contributions. The first contribution concerns the standardization of project milestones within GES Source, Amazon's internal project tracking platform. Through a Delphi-inspired collaborative process involving Program Managers across multiple regions, standardized templates were developed for projects of different sizes (XS, S, M, and L) covering both Capacity and Non-Capacity initiatives.