This paper presents a novel approach for streamlining data collection and wrangling processes to facilitate Machine Learning (ML) and Artificial Intelligence (AI) applications in generating Level of Detail 2+ (LoD2+) urban models. In the era of rapid urbanization, accurate and dynamic 3D city modeling has become indispensable for urban planning, disaster management, and Geographic Information Systems (GIS). The proposed methodology leverages the capabilities of public domain data sources like Google Street View and OpenStreetMap to assist in creating comprehensive urban digital twins. By integrating diverse datasets within a unified data model, we aim to overcome the limitations posed by traditional urban modeling techniques. The Digital Twin Cities Centre (DTCC), hosted by Chalmers University of Technology, plays a pivotal role in this endeavor, providing an open-source platform for data fusion and urban design. The work presented is a milestone towards automating LoD2+ urban digital twins creation based on non-commercial software and data sources.