Today's data science use cases, which deal with the application of machine learning, require a sufficient amount of data. The available data usually show a great variety in terms of data sources, formats and quality. Therefore, data consumers need ways to discover, understand and access potentially relevant data sources. At the same time, data providers require options to make their data sources available regardless of their later use. For both data provision and consumption, it has to be ensured that the effort that arises during these processes, called time-to-analytics, is kept low. An approach to simplify the provision and consumption of data and thereby decrease the time-to-analytics is offered by Ontology-based Data Management (OBDM). By centralizing data storage and access as well as by defining a common shared conceptualization in the form of ontologies, OBDM enables the seamless collection, integration, discovery, understanding and access of heterogeneous data sources. The disadvantage of current OBDM approaches is that they require an ontology that was created in advance. This thesis presents a novel approach, called Bottom-up Knowledge Graph (BUKG), that improves OBDM by overcoming issues of traditional ontology engineering for the management of (semi-)structured data sources. Instead of designing and maintaining ontologies top-down, the presented approach learns the individual conceptualizations of data providers and consumers and continuously integrates them into a common shared conceptualization. All in all, the presented approach makes an important contribution to the semantic data management by supporting the seamless collection, integration, discovery, understanding and access of heterogeneous data sources based on a novel bottom-up conceptualization.weiterlesen