Exploiting FAIR Data & Data Catalogs with LeapAnalysis: A New Approach to Virtualized Data Integration Across Federated Data Sources.
This workshop will extend the themes from the plenary talk on "FAIR Data, Data Catalogs and the Foundations of AI” to go deeper into the topics covered and show a path forward for realizing these items inside of a commercial software product, LeapAnalysis. We will discuss challenges for semantic implementation and what pharma companies should consider when FAIR-ifying their existing data sources. Using LeapAnalysis we will show a straight-forward and simplified way to use semantics (a Knowledge Graph) directly against existing data sources, without the need for migration, data integration steps, or making copies of the data. We will show how virtualization can be used over federated data sources directly by allowing data to be combined at the metadata level only – the instance data stay in their original sources – in this way attendees can learn to “embrace your data silos.”
We will demonstrate our new one-of-a-kind software product LeapAnalysis (LA), which utilizes FAIR principles and Data Catalogs to virtually connect data sources in real time. We will show how to create custom namespaces to align data sources to a common semantic set of models for greater interpretability of the metadata. We will demonstrate how LA can seamlessly connect and utilize data sources on the fly, without the need for heavy-weight ETL or integration steps, which are the norm in other search & analytics platforms. Utilizing cloud technology, we will then run SPARQL queries against several disparate non-semantic data sources directly, in real time, and resolve those queries in a single go. We will then discuss how analytics and reasoning can be utilized against those query results to provide users with unprecedented capabilities for finding and utilizing their data for a variety of applications
Areas covered in this session
Role of metadata, specifically in terms of:
- Controlled Vocabularies
- Layered Ontologies
Who should benefit
Decision makers in R&D, IT, Data Management, Analytics, Labs, Enterprise Information Management and Data Science.
Why should you attend
Join us to see the future of semantic technologies, machine learning and a real pathway to AI. Learn how to avoid many of the pitfalls of building cumbersome and ineffective data lakes and data warehouses.
Eric Little works on the research & development of semantic technologies and their application to analytics, including graph-based computing, data modeling and advanced computational reasoning within pharmaceutical and biotech verticals.
Mr. Little has extensive experience designing and building scalable semantic systems for a wide range of customers and projects across several industry verticals. He has helped many of his customers improve their abilities to structure and manage their data, allowing them to find new types of patterns in their data that link directly to important business drivers within the enterprise.
OSTHUS is a multinational data consultancy headquartered in Aachen, Germany, with data innovation frameworks deployed in 16 of the top 20 big pharmaceutical companies in the world. For two decades, we have worked with executive teams in diverse industries to introduce a culture of innovation that drives enterprise value. We offer a strong track record of driving costs down while increasing the time to market for complex data initiatives. OSTHUS Blueprints serve as the choice methodology for companies seeking to leverage a vendor-agnostic approach in the pursuit of global digitization, artificial intelligence, prescriptive and deep learning efforts.