The dbt (data build tool) and its use when building data warehouses

The thesis focuses on the use of the dbt (data build tool) for designing and implementing modern data warehouses. This tool facilitates the management of data processes, ensures data quality, supports code versioning, and enables documentation creation. The aim of the thesis is to demonstrate, based on a literature review and practical implementation, how dbt can be effectively used in designing a data warehouse and which techniques and approaches can serve as inspiration for similar projects. The thesis is divided into two parts – theoretical and practical. The theoretical part describes the key features of the dbt tool, its structure, and the main principles of its operation. It also highlights the modular approach and the possibilities of integration with modern cloud platforms such as Google BigQuery. The practical part builds on the findings of the theoretical section and applies them to a real-world project in the healthcare sector. In this part, a data warehouse is designed with an architecture divided into layers L0 to L3, with transformation processes fully managed by dbt. The output of the practical part is a proposed approach to solving data projects using dbt, which includes templates and best practices, such as the use of macros for recurring transformations or the implementation of historical approaches through snapshots. The thesis demonstrates that dbt can modernize data architecture and streamline processes associated with data warehouses, with a strong emphasis on transparency, documentation, and data validation.