2. Why BlueBrain Nexus?
• Trans-disciplinary science
• Issue: domain specific data platforms
• Need to capture and track data provenance
• FAIR principles
4. Nexus enables data modeling best practices
Reference implementation:
INCF/neuroshapes
5. Data modeling with Nexus
We’ve adopted Semantic Web Technologies to:
• describe and constrain data
• connect and relate data
• evolve and migrate the data models
• search and discover
• lens through the graph
• interoperate with systems that follow the same standards
Nexus is based on:
• JSON-LD (RDF) as the default exchange format
• W3C SHACL as a metadata constraint language
• W3C PROV-O as the standard to capture and query provenance
6. What is Nexus?
• a data management platform
• a metadata catalog
• a semantic search engine
• a FAIR publishing platform
• domain agnostic
• treats provenance as a first class citizen
• secure
• highly concurrent and scalable
• extensible
Science is becoming trans-disciplinary (e.g. Brain science, Venice Time Machine, Physics, Astronomy…)
Data integration platforms are traditionally built in domain specific manner.
Challenge: how to build a single platform to deal with issues like: multiple domains integration, distributed data, providing an integrated view to enable discovery, analysis, AI applications…
Data driven science, involves knowledge discovery which requires tracking of data > assess quality > build trust > enable reproducibility all of which are enabled by provenance.
With the rise of team science (solving problems that leverage the expertise of professional from different fields), it is more important than ever to play by the sociological drivers that enable science, and attribution/credit work is a key part.
FAIR principles address a lot of these concerns but not giving solution.
We need trans-disciplinary data integration, this is why we have built Blue Brain Nexus, to implement the FAIR principles and support data driven open science by enabling communities to develop their own vocabularies to describe data, supporting multi-domain modelling, deep integration of provenance …
We offer a platform that is scalable, deeply engineered, based on open standards to support cross domain, provenance tracked, FAIR data that is born connected.
Add link to INCF Neuroshapes as a reference implementation of the data modeling best practices for neuroscience
SHACL: expressiveness, interoperable, processable (computers can make sense of it), domain agnostic