Public UBS Knowledge Graph – Building a Connected Data Catalog

In this podcast episode, Gregor Wobbe, Head of Data Architecture at UBS, shares insights into the bank's innovative approach to large-scale data integration and management. He highlights the complexities of overseeing data across 5,000 business applications, each with its unique database and structure. Wobbe introduces a new federated strategy that utilizes data services, specifically Kubernetes virtual machines, to convert data into JSON-LD linked data, which is then processed through Kafka and stored in both a graph database and a data lake. This strategy enhances data discoverability, comprehension, trust, utilization, and reuse by employing open standards like DCAT, DQV, and PROV. The main takeaway is UBS's shift from a centralized, tool-centric model to a decentralized, capability-driven approach, leading to better scalability and simpler data integration.