GraFlex - Interactive Graph Database Evolution and Graph Data Transformation

• Applicant: Professorin Dr.-Ing. Meike Klettke
• Subject Area: Data Management, Data-Intensive Systems, Computer Science Methods in Business Informatics
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 552691270

If graph databases are used as the primary format for data storage, then the entire ecosystem for graph data processing is required. In particular, methods for graph schema evolution and transformation of graph data are currently not yet available. In this project, we want to develop solutions for these two tasks. The project aims to create an evolution component for graph databases that performs schema changes in graph databases in a similar way to the “alter table” statements of relational databases. In graph databases, evolution includes both the implicit structural changes within the graph data and the changes of their explicit (optional) schemas. The approach to be developed will be designed interactively from the outset. This means that, compared to the usual batch processes of evolution, a human-in-the-loop approach is developed here, in which users can make informed decisions and thus determine and control the evolution. To enable this, the approach will also visualize a pre-estimation of the changes caused by evolution operations and develop a preview component for graph schema evolution. The method will then be extended to general graph data transformations. These are needed, for example, if certain standards are used in a scientific discipline or if graph databases are to be integrated into an existing data format. For this purpose, the project will first develop a method for matching source and target schema, from which general transformation operations (SMOs) will then be derived. In the final step, these SMOs will be used to transform the graph database into the target structure. As with schema evolution, this process will be designed as an interactive process from the outset so that it will be easy to be used by computer scientists and application experts. For both tasks (evolution and transformation of graph databases), we want to conceptually develop methods that are not yet available and implement them in prototypical tools.

• DFG Programme: Research Grants