Modern underground infrastructure projects generate immense streams of real-time geomechanical data during the active construction phase. However, high-value field records, such as geological face mappings, typically suffer from severe format heterogeneity, remaining trapped within unstructured PDF reports and isolated spreadsheets. This technical fragmentation creates operational data silos, forcing engineering teams into slow, error-prone manual data transcription that delays critical updates to 3D geological synthesis models. To bridge this technological bottleneck, this thesis designs, implements, and validates an automated, data-centric management methodology applied as a Proof of Concept (PoC) within the Tunnel Euralpin Lyon Turin (TELT) project. The developed data pipeline consists of two primary automation phases.