Data Management and Modeling in Motorsport Racing

Motorsport teams increasingly rely on data to optimize vehicle performance, improve reliability, and support strategic decisions. For student motorbike racing teams, telemetry, setup and inventory data are generated continuously but often remain fragmented, inconsistently managed and difficult to analyze due to resource constraints and heterogeneous formats. This thesis addresses the challenge of designing and implementing an effective data management solution that enables reliable storage, integration and retrieval of such data. The research follows a structured methodology. First, requirements were elicited through collaboration with team stakeholders, identifying functional needs (e.g.\ easy telemetry access, comparative setup analysis, inventory tracking) and non-functional constraints (low operational overhead, maintainability, cost efficiency). Second, candidate data management architectures were studied, comparing relational and non-relational databases, data warehouse and data lake and lakehouse patterns. Third, an optimal hybrid architecture was proposed and implemented. Finally, the prototype was validated with real-world team data, qualitatively assessing ingestion reliability, query performance, data quality and usability. Results show that the proposed architecture consolidates heterogeneous data, reduces manual effort in data preparation, and enables faster and more reliable analyses through curated schemas and dashboard interfaces. While limitations remain—particularly in automated data-quality handling, large-scale telemetry archiving, and data lineage —the solution demonstrates that a pragmatic, low-cost approach can deliver substantial benefits to student motorsport teams. The thesis contributes both a working prototype and a generalizable methodological framework that other teams can adopt or extend, laying the groundwork for future enhancements such as predictive modeling, advanced analytics and scalable long-term storage.