The company's platform leverages public open data sources for marketing campaign optimisation. The effectiveness of such a campaign depends crucially on external factors such as political opinion, weather, surrounding demographics and even district-level trends. Research questions are: How can data sources be classified, and thus the integration be sped up, considering that often practical restrictions on anonymous access, request limits and heterogeneous data formats exist? How can multiple data sources be logically bundled, e.g. citizen data from multiple cities when looking for just “citizen information“? Can regression testing and alerting be implemented in a generic way to avoid the negative influences in sudden data changes due to network outages or invalidated URLs?