Research Graph Schema

Research Objects

Research Graph is composed of five primary nodes, also known as research objects:

Researcher: An individual who conducts scientific or scholarly study, often within a specific field of knowledge.
Publications: Academic publications include journal articles, books, patents, conference abstracts, and theses, serving as the primary medium for disseminating research findings.
Research Data: Data that is used or generated in the course of academic studies or research project.
Grants: Awards and grants are financial support provided by funding bodies (like government agencies, foundations, or institutions) to researchers or research projects.
Organisations: This encompasses institutions like universities, research institutes, and academic societies.

The following tables describe the metadata (propeties, types and constraints) for each node (research object). We have also provided an example of the JSON representation of each research object.

Researcher

Property	Type	Description
`key` ◆	`string`	Unique URI for research object identification 📐 Format: `@source/@local_id`
`source` ◆	`string`	Metadata source identifier 💡 Use domain name without extension 📝 Examples: `orcid`, `datacite`, `nih`
`local_id` ◆	`string`	Record identifier in source database
`last_updated` ◆	`dateTime`	Update timestamp in Research Graph
`full_name` ◆	`string`	Complete name when parts unavailable ℹ️ Fallback for missing first/last names
`first_name`	`string`	First name (English preferred)
`last_name`	`string`	Last name (English preferred)
`url`	`string`	Resource URL
`orcid`	`string`	ORCID ID (`xxxx-xxxx-xxxx-xxxx`) 🔗 Structure documentation
`scopus_author_id`	`string`	Scopus author identifier

◆ Required field

Example Record

{
  "key": "orcid/0000-0002-4259-9774",
  "source": "orcid",
  "local_id": "0000-0002-4259-9774",
  "last_updated": "2024-03-15T14:30:00Z",
  "full_name": "Amir Aryani",
  "first_name": "Amir",
  "last_name": "Aryani",
  "url": "https://orcid.org/0000-0002-4259-9774",
  "orcid": "0000-0002-4259-9774",
  "scopus_author_id": "35068996400"
}

Publication

Property	Type	Description
`key` ◆	`string`	Unique URI for research object identification 📐 Format: `researchgraph.ai/@source/@local_id`
`source` ◆	`string`	Metadata source identifier 💡 Use domain name without extension 📝 Examples: `orcid`, `datacite`, `nih`
`local_id` ◆	`string`	Record identifier in source database
`last_updated` ◆	`dateTime`	Update timestamp in Research Graph database
`title` ◆	`string`	Publication title
`author_list` ◆	`string`	Comma-separated list of co-authors ℹ️ Complements title for easy recognition
`publication_type`	`string`	Publication type 🔗 ORCID work types
`url`	`string`	Resource URL
`doi`	`string`	Digital Object Identifier 📐 Format: `doi_prefix/doi_suffix`
`isbn`	`string`	International Standard Book Number
`publication_year`	`integer`	Publication year (YYYY format)
`scopus_eid`	`integer`	Scopus EID Unique academic work identifier in Scopus

◆ Required field

Example Record

{
 "key": "crossref/10.1038/sdata.2018.99",
 "source": "crossref",
 "local_id": "10.1038/sdata.2018.99",
 "last_updated": "2024-03-15T14:30:00Z",
 "title": "A Research Graph dataset for connecting research data repositories using RD-Switchboard",
 "author_list": "Aryani, A., Poblet, M., Unsworth, K., Wang, J., Evans, B., Devaraju, A., Hausstein, B., Klas, C.-P., Zapilko, B., Kaplun, S.",
 "publication_type": "journal-article",
 "url": "https://doi.org/10.1038/sdata.2018.99",
 "doi": "10.1038/sdata.2018.99",
 "publication_year": 2018
}

Dataset

Property	Type	Description
`key` ◆	`string`	Unique URI for research object identification 📐 Format: `researchgraph.ai/@source/@local_id`
`source` ◆	`string`	Metadata source identifier 💡 Use domain name without extension 📝 Examples: `orcid`, `datacite`, `nih`
`local_id` ◆	`string`	Record identifier in source database
`last_updated` ◆	`dateTime`	Update timestamp in Research Graph database
`title` ◆	`string`	Dataset title
`author_list` ◆	`string`	Comma-separated list of creators or curators ℹ️ Complements title for easy recognition
`url`	`string`	Resource URL
`doi`	`string`	Digital Object Identifier 📐 Format: `doi_prefix/doi_suffix`
`publication_year`	`integer`	Publication year (YYYY format)

◆ Required field

Example Record

{
 "key": "zenodo/10.5281/zenodo.4939953",
 "source": "zenodo",
 "local_id": "10.5281/zenodo.4939953",
 "last_updated": "2024-03-15T14:30:00Z",
 "title": "People with disability participated in labour force 2018",
 "author_list": "Woo, J., Aryani, A.",
 "url": "https://doi.org/10.5281/zenodo.4939953",
 "doi": "10.5281/zenodo.4939953",
 "publication_year": 2021
}

Grant

Property	Type	Description
`key` ◆	`string`	Unique URI for research object identification 📐 Format: `@source/@local_id`
`source` ◆	`string`	Metadata source identifier 💡 Use domain name without extension 📝 Examples: `orcid`, `datacite`, `nih`
`local_id` ◆	`string`	Record identifier in source database 💡 Best practice: `grant_number`
`last_updated` ◆	`dateTime`	Update timestamp in Research Graph database
`title` ◆	`string`	Grant title
`url`	`string`	Resource URL
`purl`	`string`	Persistent URL 📝 Example: `http://purl.org/au-research/grants/arc/SR0567397`
`doi`	`string`	Digital Object Identifier 📐 Format: `doi_prefix/doi_suffix`
`publication_year`	`integer`	Publication year (YYYY format)
`funder`	`string`	Funding organization 💡 Best practice: funder domain address 📝 Example: `arc.gov.au`
`funding_amount`	`long`	Funding amount 📝 Example: `450000`
`funding_currency`	`string`	Currency code 💡 Use ISO 4217 codes 🔗 ISO 4217 reference
`start_year`	`integer`	Grant start year (YYYY)
`end_year`	`integer`	Grant end year (YYYY)

◆ Required field

Example Record

{
 "key": "arc/DP210103512",
 "source": "arc",
 "local_id": "DP210103512",
 "last_updated": "2024-03-15T14:30:00Z",
 "title": "Advanced Machine Learning Techniques for Climate Change Prediction",
 "url": "https://dataportal.arc.gov.au/NCGP/Web/Grant/Grant/DP210103512",
 "purl": "http://purl.org/au-research/grants/arc/DP210103512",
 "doi": "10.13039/501100000923",
 "publication_year": 2021,
 "funder": "arc.gov.au",
 "funding_amount": 875000,
 "funding_currency": "AUD",
 "start_year": 2021,
 "end_year": 2024
}

Organisation

Property	Type	Description
`key` ◆	`string`	Unique URI for research object identification 📐 Format: `@source/@local_id`
`source` ◆	`string`	Metadata source identifier 💡 Use domain name without extension 📝 Examples: `orcid`, `datacite`, `nih`
`local_id` ◆	`string`	Record identifier in source database
`last_updated` ◆	`dateTime`	Update timestamp in Research Graph database
`name` ◆	`string`	Organisation name 💡 Best practice: use English label from WikiData records
`url`	`string`	Organisation URL 📝 Example: `http://www.monash.edu/`
`grid`	`string`	GRID identifier 📝 Example: `grid.1002.3`
`doi`	`string`	Digital Object Identifier 📐 Format: `doi_prefix/doi_suffix`
`ror`	`string`	ROR identifier 📝 Example: `02bfwt286`
`isni`	`string`	ISNI identifier 📝 Example: `0000000419367857`
`wikidata`	`string`	Wikidata identifier 📝 Example: `Q598841`
`country`	`string`	Country 📝 Example: `Australia`
`city`	`string`	City 📝 Example: `Melbourne`
`latitude`	`float`	Latitude coordinate 📝 Example: `-37.908333333333`
`longitude`	`float`	Longitude coordinate 📝 Example: `145.13805555556`

◆ Required field

Example Record

{
 "key": "ror/02bfwt286",
 "source": "ror",
 "local_id": "02bfwt286",
 "last_updated": "2024-03-15T14:30:00Z",
 "name": "Monash University",
 "url": "http://www.monash.edu/",
 "grid": "grid.1002.3",
 "doi": "10.13039/501100001779",
 "ror": "02bfwt286",
 "isni": "0000000419367857",
 "wikidata": "Q598841",
 "country": "Australia",
 "city": "Melbourne",
 "latitude": -37.908333333333,
 "longitude": 145.13805555556
}

Design Principles

The Research Graph schema is designed to enable transformation of scholarly information to accessible and interoperable networks. The schema follows the following design principles:

East of use: The schema enables the graph databases to be easy to collect, connect, and work with by software developers.
Highly interoperable with PID systems: The schema enables Research Graph content to be early linked to PID systems such as ORCID, Crossref, DataCite, and other providers of persistent identifiers. Also, it enables the Research Graph to be easily translated to other schemas using the PID ecosystem.
Highly computable: The schema enables Research Graph content to be light and scalable, providing the foundation for storing and computing very large graphs using only personal computers instead of high-performance computing.

Aligned with these principles, the schema is not comprehensive and does not include a complex and complete set of metadata for all possible data sources. Instead, it encourages the Research Graph systems to use external databases to augment the graph data when more comprehensive analysis is needed.