Interactive Information Visualization allows users to analyze abstract datasets through graphical representations, providing significant usability advantages over tabular approaches. Although some graphical toolkits for the creation of visualizations exist, they in some way lack the expressiveness of text-based frameworks. While datasets are always specific to one use case, users are therefore often confined to fixed sets of visualizations with limited customization options, as the creation of custom solutions is still an expert task.\nIn the realm of software development, Model-Driven Engineering allows the creation of domain-specific graphical editors that eliminate the need to manually write code to create software. Hence, no programming experience is needed, enabling domain-experts themselves rather than software developers to work on a project. Additionally, existing graphical solutions can be reapplied to multiple use cases and easily extended with new features.\nThis thesis sets out to apply the techniques of Model-Driven Engineering to the domain of Interactive Information Visualizations. A survey on existing approaches to the creation process serves as the basis for a set of design concepts, which fulfill common requirements in this domain. Afterwards, a first prototypical implementation of these concepts is evaluated in a set of case studies in terms of functional flexibility as well as the modularity and extensibility of its components.