Extract, Transform, Load (ETL) processes are an essential part of building data warehouses. They allow data from many different systems to be collected, cleaned, and stored in a form that can be used for analysis. In Microsoft SQL Server Integration Services (SSIS), these processes are usually created by manually configuring packages. Although this approach provides flexibility, it is repetitive, slow and prone to errors, especially when the system must handle a large number of sources or when several developers are working on the same project. This thesis proposes a metadata driven automation framework written in Python to simplify this task. Instead of separately build each SSIS package by hand, the system utilizes metadata stored in structured, developed by hand, Excel files and automatically generates packages in XML format.