Data dictionary cookbook for research data and software interoperability at global scale

We are now facing profound changes (biodiversity, climate, pandemic, etc.). Human impacts and their mitigation will depend on our ability to mobilize research at the global level. The sustainable development of the society will largely depend on the sustainable development of global science and scientific research tools, outputs, and research ecosystems. This globalization of research requires interoperating our observation and experimentation systems in order to better understand these changes, to better simulate their effects.

The Covid-19 pandemic is now raging around the world. The reproducibility of research and results across regions in different contexts should accelerate human responses. Data sharing and the development of Synthesis Research with data aggregation at large scale is critical to enable such processes. The use of common knowledge, vocabularies, standards and procedures at a large scale is necessary.

The objective of this poster is to report on the challenges met while building data dictionaries in three global projects related to biodiversity and/or disease research: PARSEC, Kakila, ERINHA-Advance. The Kakila database centralizes and harmonizes marine mammal observation data for the AGOA sanctuary around the French archipelago of Guadeloupe, French Antilles. The PARSEC Project is building new tools for data sharing and reuse through a transnational investigation of the socioeconomic impact of protected areas. The ERINHA-Advance project aims to support the operations of the ERINHA research infrastructure which is designed to generate data from transnational access research activities on highly pathogenic agents. In these 3 global case-studies, similar challenges have arisen: to aggregate and interoperate pre-existing heterogeneous data at the global scale, and to share common tools to monitor, maintain quality,  scan scale and cope with uncertainty. This poster proposes a draft common methodology, a  data dictionary cookbook, which will provide a roadmap towards the building of large scale - data dictionaries. Topics proposed to be covered in such a cookbook include:

• how to search for existing and appropriate data dictionaries, controlled vocabularies or other semantic resources (before building a new one),

• the first steps for data dictionary building,

• data dictionary literacy (and why it is a mandatory work),

• how to define all scientific objects, aspects (or use existing one) and agree on the definitions with the whole community,

• building / proposing variables / indicators with ontology models, schemas, variables naming rules and context awareness,

• and finally addressing dimension issues considering each context.

The common experience of our three projects showed that we need to proceed step by step as simply as possible and to ensure that each step is understandable for the whole community. It is necessary to improve access and re-use of all existing semantic materials and not trying to build a cathedral with a little spoon.