Heterogeneous data to knowledge graphs matching

Many applications rely on the existence of reusable data. The FAIR (Findability, Accessibility, Interoperability, and Reusability) principles identify detailed descriptions of data and metadata as the core ingredients for achieving reusability. However, creating descriptive data requires massive manual effort. One way to ensure that data is reusable is by integrating it into Knowledge Graphs (KGs). The semantic foundation of these graphs provides the necessary description for reuse. In the Open Research KG, they propose to model artifacts of scientific endeavors, including publications and their key messages. Datasets supporting these publications are essential carriers of scientific knowledge and should be included in KGs. We focus on biodiversity research as an example domain to develop and evaluate our approach. Biodiversity is the assortment of life on earth covering evolutionary, ecological, biological, and social forms. Understanding such a domain and its mechanisms is essential to preserving this vital foundation of human well-being. It is imperative to monitor the current state of biodiversity and its change over time and to understand its forces driving and preserving life in all its variety and richness. This need has resulted in numerous works being published in this field. For example, a large amount of tabular data (datasets), textual data (publications), and metadata (e.g., dataset description) have been generated. So, it is a data-rich domain with an exceptionally high need for data reuse. Managing and integrating these heterogeneous data of biodiversity research remains a big challenge. Our core research problem is how to enable the reusability of tabular data, which is one aspect of the FAIR data principles. In this thesis, we provide answer for this research problem.