One important characteristic of Big Data and modern analytics in general is its structural variety. Modern data analytics involve data in very different data models, such as semi-structured JSON or XML data, relational data, and graph data. Typical database management systems implement a single data model, which makes it difficult to handle multi-model analytical tasks. In the future, database management systems will have to support multiple data models and processing such data in a federated process, i.e., data from different data models can be processed together in a single query or a single transaction. A challenge is that modern datasets lack the structure of relational data and are hence more difficult to optimize for management and querying. Essential help can be given to these users by interactive and visual components that guide them through the query formulation process. Likewise, result presentation is more difficult for modern datasets, and presenting results of graph data in tables can easily become counter-intuitive in some applications. This project aims to enable analysis over modern data sets. To enable this, various aspects have to be addressed, for instance the mapping of data between different models to answer a single query, the integration of different query languages, intuitive presentation of results, storage layout, and the efficient execution of queries.

Main Advisor at Aalborg Universitet (AAU)
Co‐advisor at Universitat Politècnica de Catalunya (UPC)