With increasing popularity of interactive data analysis, simulations and visualization, there is a need for flexibility in the development and use of data management toolkits. Although some work has done by the researchers, a solid foundation is not yet established. This thesis introduces Datavis, a flexible, extensible data management toolkit that integrates simulation systems into data visualization systems. Datavis links together standard numerical simulation tools such as EASY5 [Bl] with visualization packages like GGobi making the data flow among these different tools transparent to the user. Datavis allows users to perform data analysis interactively independent of the data format or the packages needed in the process. Datavis has three main components: The core that handles the user requests, the data handler which takes care of the organization and storage of the data, and the visualization handler which takes care of the instantiation and manipulation of the visualization system. The proposed toolkit has been successfully implemented and transferred to use as part of routine data analysis activities at a company. At the same time, the work on Datavis has laid the foundation for more sophisticated approaches on integrating simulation and visualization for data analysis.