1 of 24

Developing a FAIR-Compliant Metadata Schema for Experimental Nuclear Physics

Ivan Knežević, Dr Andrew K. Mistry GSI Helmholtzzentrum für Schwerionenforschung

CC-BY 4.0

GSI Helmholtzzentrum für Schwerionenforschung GmbH

2 of 24

GSI Helmholtz Centre for Heavy-ion Research

2

Who we are?

Founded in 1969
Elements H to U can be accelerated

Past Achievements

Discovering six new elements, many exotic nuclei
Developing a new cancer treatment

Our future

First stage injector for the FAIR facility
Experiments continuing (FAIR Phase-0)

Research Interests

Accelerator Physics
Detector Development
Atomic Physics
Nuclear and particle Physics
Plasma Physics
Biophysics
Materials Research
High-Performance Computing
Theoretical Physics

Heavy-ion accelerator laboratory in Darmstadt, Germany with 1350+ employees & 700 external users

GSI Helmholtzzentrum für Schwerionenforschung GmbH

3 of 24

FAIR (Facility for Antiproton and Ion Research in Europe)

3

GSI

FAIR

New accelerator facility

Top priority for European Nuclear Physics Community
International: 50 countries, 3000 researchers
Diverse community from atomic to particle physics
‘FAIR goes F.A.I.R.’: commitment to open science
Towards the next generation “data challenge” Volume, Velocity, Veracity, Variety, and Complexity
~TB/s data rates, online processing, ~5x105 cores
Data stored to disk 40+ PB/year
Distributed computing with a large user community
Data preservation and accessibility key to success

GSI Helmholtzzentrum für Schwerionenforschung GmbH

4 of 24

Research Data and Software at GSI: Rich and varied

4

Results

Raw data

Pre-processed data

Research Area	Exp. run time	Raw data	Simulations	Final datasize
Materials science	~Minutes	~MB-GB	-	10MB
HADES	3 weeks	130TB	150TB	<1TB
FAIR CBM	2 months	18PB	9PB	4PB

GSI Helmholtzzentrum für Schwerionenforschung GmbH

5 of 24

F.A.I.R. data principles and metadata

Metadata is essential in Research Data Management and Open Science to enable FAIR data (and code)

5

Icons made by Freepik are licensed by CC BY 3.0

FAIR resources graphic is licensed under a CC 4.0 International License

Allows datasets to be searched for and found
Enables interoperability between datasets
Facilitates data reprocessing
Promotes efficient use of resources

GSI Helmholtzzentrum für Schwerionenforschung GmbH

Let's begin by examining the FAIR data principles, which serve as a foundational framework for effective data management and open science.

Findable: Data should be easy to find for both humans and computers. This means assigning unique identifiers and providing rich metadata.

Accessible: Once found, data should be retrievable using standardized protocols, potentially with authentication and authorization where necessary.

Interoperable: Data should be compatible with other data and tools. This involves using standardized formats and vocabularies.

Reusable: Data should be well-described so it can be replicated and/or combined in different settings. Clear usage licenses and provenance information are key.

As we can see, metadata is the common thread that weaves through all these principles. It's the descriptive information that makes data findable, accessible, interoperable, and reusable.

Metadata acts as the roadmap to our data assets. Without it, datasets are like books without titles or content pages—difficult to discover and utilize.
By providing detailed descriptions, metadata enables researchers to locate datasets relevant to their work, enhancing discoverability.
Standardized metadata ensures that data from different sources can be integrated and understood collectively, which is vital for collaborative research.
Comprehensive metadata allows others to understand the context, methodology, and parameters of data collection, making it possible to replicate studies or apply data in new analyses.
By making data easily accessible and understandable, we reduce duplication of effort and accelerate scientific discovery.

6 of 24

Underlying problems in Data Management

Growing data volumes

Unstructured data

Reproducibility crisis

Lack of user awareness in RDM practices

6

Exabytes (Billions of Gb)

GSI Helmholtzzentrum für Schwerionenforschung GmbH

However, despite the clear benefits of proper metadata management, we face several significant challenges in the realm of data management that impede our ability to fully realize the FAIR principles.

First, we are experiencing an unprecedented growth in data volumes. As this graph illustrates, the last 15 years have seen an explosive increase in the amount of data generated—a phenomenon often referred to as a 'Cambrian explosion' of data. This phenomenon has not bypassed scientific data either. While this wealth of data holds immense potential, it also presents significant challenges in terms of storage, management, and analysis. The sheer volume can overwhelm our current systems and practices, making it difficult to keep track of data and extract meaningful insights.

Second, much of this data is unstructured. This segment of the graph highlights the proportion of data that is unstructured—data that doesn't fit neatly into traditional databases or formats. Unstructured data is inherently more difficult to organize, search, and analyze. Without proper structure and metadata, valuable data can become lost in the noise, inaccessible to those who could benefit from it.

Third, we are facing a reproducibility crisis in science. According to a well-known survey published in Nature in 2016, approximately 40% of researchers reported that they could not reproduce even their own experiments. Factors contributing to this crisis include inadequate documentation, lack of access to original data, and insufficient metadata.

7 of 24

Standard Workflow in Experimental Nuclear Physics

Workflow is representative of experimental procedures at GSI and similar facilities.

Each stage generates specific metadata crucial for data management.

Our metadata stratification and categorization is grounded in this workflow.

Definition of Digital Research Product (PUNCH4NFDI) - https://www.punch4nfdi.de/

7

GSI Helmholtzzentrum für Schwerionenforschung GmbH

Having explored the significant challenges we face in data management, it's clear that we need a systematic approach to manage our research data effectively. Let's take a closer look at the standard workflow in experimental nuclear physics, particularly at GSI, to understand where metadata plays a crucial role in addressing these issues. Here, we have a diagram representing the standard workflow for conducting experiments at GSI, which is quite representative of procedures at similar experimental facilities. This workflow encompasses all the critical stages of a research project, from the initial proposal to the final publication.

Proposal Management System: Leads to Data Management Plan in RDMO.

Experiment Execution: Data acquisition phase.

Data Archiving: Long-term storage solutions.

Parallel Distributed File System: For immediate data access.

Data Analysis: Utilizing HPC cluster computing infrastructure.

Code Repository: Integration with GSI GitLab.

Publication: Final research outputs.

Our metadata stratification and categorization are grounded in this workflow. By aligning metadata types with each stage, we can create a comprehensive and organized metadata schema.

This approach aligns with the concept of Digital Research Products (DRPs), as defined by the PUNCH4NFDI initiative. DRPs encompass all digital outputs of the research process.

8 of 24

Digital Research Product

8

Infrastructure

Dataset

Publication

Software

Workflow

REANA, Compute Resources, Containers�ENV, distors, ContainerID

Infrastructure

Accelerators, Detectors, Radio telescopes instrument records�Versioning, DOI

Dataset

Zenodo, Public and Collaboration data repositories�DOI, PID

Publication

arXiv, journals, ADS, open access�DOI, ISBN

Software

GitHub, GitLab, ROOT, CASA�Versioning DOI

“All digital outputs of the research process have to be referenceable”

Project

GSI Helmholtzzentrum für Schwerionenforschung GmbH

9 of 24

Project Overview

Addressing problems:

Growing unstructured data,
Diverse data formats and nomenclatures,
Difficulty in data sharing across institutions,
No common schema yet between nuclear physics experiments.

9

By Randall Munroe: https://xkcd.com/927

Objectives:

Develop a standardized, adaptable metadata schema.
Facilitate data input, referencing, management and publication.

Nuclear, Astro, and Particle Metadata Integration for eXperiments

GSI Helmholtzzentrum für Schwerionenforschung GmbH

10 of 24

Example: GSI accelerator facility

10

Photo: © GSI: G. Otto

Photo: © GSI J. Hosan

Photo: T. Ernsting/HA Hessen Agentur

Ion sources

Type,

Projectile Ion…

Accelerators

Beam Energy

Beam Intensity

Separators

Type, Reaction…

Experimental areas

Produced Isotope

Experimental setup…

Photo: © GSI: G. Otto

Unilac

SIS18

FRS

ESR

HPC

Green Cube

Dataset location, size, code repository…

GSI Helmholtzzentrum für Schwerionenforschung GmbH

11 of 24

Importance of Standardized Metadata

Issues with non-standardized metadata,
Impact on data discovery and utilization,

11

Standardization facilitates automation and machine readability,
Facilitates data sharing and collaboration
Enhances data quality and integrity

GSI Helmholtzzentrum für Schwerionenforschung GmbH

12 of 24

Metadata stratification

12

Experiment specific metadata

Reaction data, gamma-ray energies, time of flight parameters, instrument records

3

Discipline specific metadata

Projectile data, target data, infrastructure, detector type, datasets, software repositories

2

Project specific metadata

Project name, publication date, collaborations, researchers, institutions, journal references

1

GSI Helmholtzzentrum für Schwerionenforschung GmbH

13 of 24

Metadata stratification

13

Findable

Accessible

Metadata level

Discipline Specific: Projectile, target, detector type, datasets…

Experiment Specific: Gamma-ray energies, Time of flight parameters…

Project Specific: Standard project information, names, dates, institution…

Interoperable

Reusable

GSI Helmholtzzentrum für Schwerionenforschung GmbH

14 of 24

The Metadata Schema Design

Nodal, Multi-Layered Structure:

Keeps data in common on top
Clusters separated but interconnected

14

Project

Cluster

Nuclear physics experiment

cluster

Dataset Cluster

Astrophysics experiment

cluster

Particle physics experiment

cluster

Metadata Enrichment:

New tables or fields can be added to the Experiment cluster to capture additional nuclear physics metadata,�
Other domains can create their own specific clusters while sharing the common Project and Dataset layers.

GSI Helmholtzzentrum für Schwerionenforschung GmbH

15 of 24

The Metadata Schema Design

Nodal, Multi-Layered Structure:

Keeps data in common on top
Clusters separated but interconnected

15

Project

Cluster

Nuclear physics experiment

cluster

Dataset Cluster

Astrophysics experiment

cluster

Particle physics experiment

cluster

Benefits:

Customization: Researchers can capture all necessary metadata specific to their experiments,
Findability: Clearly defined relations between the clusters makes search easier
Interoperability: The common layers ensure that data remains interoperable with other systems and disciplines,
Collaboration: Facilitates collaboration by allowing different domains to understand and use each other's data.

GSI Helmholtzzentrum für Schwerionenforschung GmbH

16 of 24

Implementation process

16

Implementation

Concept

Design

GSI Helmholtzzentrum für Schwerionenforschung GmbH

17 of 24

The Database Schema Design

17

Experiment cluster

GSI Helmholtzzentrum für Schwerionenforschung GmbH

18 of 24

Metadata Schema Generator

A tool that allows users to export comprehensive metadata for a selected project.
Supports multiple formats and encoding standards.

18

GSI Helmholtzzentrum für Schwerionenforschung GmbH

19 of 24

Metadata Schema Generator

19

JSON schema

XML schema

GSI Helmholtzzentrum für Schwerionenforschung GmbH

20 of 24

Metadata Schema Generator - data serialization

Serializing each smaller cluster separately,
Increased schema flexibility,
Relationships maintained

20

GSI Helmholtzzentrum für Schwerionenforschung GmbH

21 of 24

Metadata Schema Generator - Custom attribute names

Attribute names can be customized via code (array mapping),
Supports interoperability across platforms and institutions,
Provides flexibility to meet specific metadata standards,
Plans for making attribute customization more accessible through the UI.

21

Original attribute name

Desired attribute name

GSI Helmholtzzentrum für Schwerionenforschung GmbH

22 of 24

Interoperability with Other Systems

Designed to integrate with research data management systems,
Outputs fully compliant with international metadata standards like DataCite,
Can be used when depositing data in repositories like Zenodo,
Planned support for standard data exchange protocols like OAI-PMH*,

Easier harvesting and synchronizing metadata between the systems.

*Open Archives Initiative Protocol for Metadata Harvesting

22

Sören Fleischer, Rouven Spreckels

GSI Helmholtzzentrum für Schwerionenforschung GmbH

23 of 24

Collaboration and Synergies

Open Science Initiatives:

Funded by:

23

Involved Institutions:

GSI Helmholtzzentrum für Schwerionenforschung GmbH

24 of 24

Future Outlook

Frontend development

User Interface,
Advanced search filters,

Additional features

Automatic metadata import,
Attribute customization,
Collaboration mechanism,

Integration with other systems

Integration within the PUNCH4NFDI ecosystem,
Support for protocol OAI-PMH*.

Community expansion

The service will also be onboarded to the ESCAPE Open Source Software Repository (OSSR),
Inclusion to the in the EOSC marketplace,
Full accessibility to the wider scientific community.

24

Questions?

Contact us

open-science@gsi.de

*Open Archives Initiative Protocol for Metadata Harvesting

GSI Helmholtzzentrum für Schwerionenforschung GmbH