Today, the ad-hoc processing and integration of data is an important issue due to fast growing IT systems and an increased connectivity of the corresponding data sources. The overall goal is deriving high-level information based on a huge amount of low-level data. However, an increasing amount of data leads to high complexity and many technical challenges. Especially non-IT expert users are overburdened with highly complex solutions such as Extract-Transform-Load processes. To tackle these issues, we need a means to abstract from technical details and provide a flexible execution of data processing and integration scenarios. In this paper, we present an approach for modeling and pattern-based execution of data mashups based on Mashup Plans, a domain-specific mashup model that has been introduced in previous work. This non-executable model can be mapped onto different executable ones depending on the use case scenario. The concepts introduced in this paper were presented during the Rapid Mashup Challenge at the International Conference on Web Engineering 2015. This paper presents our approach, the scenario that was implemented for this challenge, as well as the encountered issues during its preparation.