This pages provides an overview on the policies the Dimensions team applies to the different types of metadata used for describing an organization.
Organization names
Organization names (or "labels") are used in Dimensions for multiple purposes: display, searching, and matching.
The Primary Label is mainly used for the display of organizations (e.g. how they appear in the Research Organization or Funder filters), which is reflected in our policy for defining this.
When searching for an organization in Dimensions, all labels (primary/display name and names in other languages) and acronyms are used. Labels and acronyms also play a very important role in matching affiliation strings from research output (e.g. publications, grants, patents, etc.) to the appropriate standardized / canonical Dimensions organization record.
General policies
The general policies below apply across all types of name labels in Dimensions (though due to changes in policies over time or in certain edge cases, they may not have been applied consistently to all records):
- Names should be in English. If an English translation is not available or not commonly used, or if the organization has explicitly requested for their name to be shown in their local language, we use the local language name (e,g, "Universidade de São Paulo" instead of "University of São Paulo").
- Diacritics and accented characters should be included, e.g. "Tromsø research foundation" or "Universidad Tecnológica de Tula-Tepeji".
- For educational organizations, campus / location names are omitted, unless the location is part of the official name, e.g. "University of California, San Diego" or "University of Newcastle Australia".
- The full name is used instead of an acronym (unless the acronym has been adopted in the primary name for the research organization, e.g. "IBM (United States)" or "LG (United States)"), and acronyms are omitted from names.
- If both options "The University of X" and "University of X" are valid, the variant without the definite articles is used.
- University systems should include the word "System" in the name, e.g. "Texas State University System".
- For companies - unless it is clearly part of the brand identity - the commercial suffix (e.g. LTD, PLC) is omitted, as it can often be confusing or misleading. For example, some organizations may be registered more than once as different legal entities, and attributing different sources of data (e.g. grants and journal articles) to the appropriate legal entities would be extremely time consuming and unnecessary.
- For multinational companies, we distinguish between entities in different political regions (e.g. countries) and include a country name in parentheses for each, e.g. "Google (United Kingdom)", "Google (Ireland)", etc.
Primary Label
As the Primary Label is mainly used for display purposes, it isn't always appropriate to choose the full legal name of an institute because that might not be the most recognisable for users. Further, many types of organizations (including universities, private companies and hospitals) have a number of legal entities registered that a user would not expect to see in the database and may find confusing to distinguish between.
Therefore, we define an official name as the name displayed on the organization's website (either in the title, the contact or "about" pages, or in the copyright details), Wikipedia page, or author affiliations. As a result, the Primary Label of an organization record is based on the official or most recognisable version of the organization’s name.
Language Labels
Language Labels displays versions of the organization name in one or more relevant languages other than that used in the Primary Label field, with a corresponding language tag. For example, for the University of Quebec, the Primary Label / display name is “University of Quebec”; the Language Label for French is “Université du Québec”.
Aliases
The Alias attribute is used for any additional or alternate names for the organization. These mainly include historical and popularly used names that do not correspond to the official name / Primary Label. For example, an Alias for the "California Institute of Technology" (Primary Label) is “Caltech”.
Acronyms
The Acronym field can store one or more official acronyms or initialisms for the organization, usually the first letters of the words from the Primary Label / display name. For example, the Acronym for the "State University of New York" (Primary Label) is “SUNY”.
Relationships between organization records
Relationships describe associations between organization records. We currently model two types of relationships:
- a relationship that defines a subordinate association, called a "parent-child relationship", and
- a relationship that describes other associations, called a "related relationship".
Subordinate (parent-child) relationships
A subordinate relationship is used to describe an association that has a clear sense of ownership, like a parent-child association. This type of relationship is often used to describe superinstitutes and their members; example: Agency for Science, Technology and Research. Other examples include the relationship between NHS hospital trusts and their member hospitals, e.g. Guy’s and St Thomas’ NHS Foundation Trust, or between university systems and their members, e.g. the University of Alaska System.
Affiliate (related) relationships
An affiliate relationship is used when two organizations are associated, but there is no clear sense of ownership between them. It can occur, for example, between a non-profit and a university, or a university and hospital. Example: Utrecht University and its associated medical centre.
Organization types
We've originally used organization types for internal QA, and to get a general idea about the different areas covered by the organizational data in Dimensions. As a result, this has been the main focus of our curation and review work historically: accepting some level of errors as the overall trends would still be visible, and errors could be corrected when coming across a record for other purposes.
Every organization record is assigned one type only. However, as types are not mutually exclusive in reality, we assign the type that fits best, according to the selection criteria below, listed according to priority. Some organizations can be difficult to categorise as they are in a grey area between the different types.
| Type | Description | Implementation | Example |
|---|---|---|---|
| Healthcare | A health-related facility where primarily patients are treated. Includes hospitals, medical centers, health centers, treatment centers. Also includes trusts and healthcare systems. | An organised provision of healthcare as the primary function will get Healthcare assigned. | A private hospital with educational responsibilities will be categorised as Healthcare. |
| Education | An educational institution where research takes place. Can grant degrees and may have faculties, departments, and schools. | Is used for any public or private organization where the primary task is education. | A public university with a medical school is categorised as Education. |
| Archive | Repository of documents, artefacts, or specimens. Includes libraries and museums that are not part of a university. | Primary function is a research archive. | A place where information and records are stored that allows access to the collection for research purposes is categorised as Archive. |
| Facility | A building or facility dedicated to research within a specific area. Usually contains specialised equipment. Includes specialist research institutes as well as laboratories and large infrastructures such as astronomical observatories and particle accelerators. | Primary function is research facility. | An astronomical observatory that provides a library service is categorised as Facility. |
| Nonprofit | An organization that uses its surplus revenue to achieve its goals. Includes charities and other non-government research funding bodies. | A non-business entity operated for a collective, public or social benefit. | A charitable organization that carries out clinical research is categorised as Non-profit. |
| Company | A legal entity with the aim of engaging in business and gaining profit. | Primary function is profit-making business. | An organization that sells healthcare products and conducts research is categorised as Company. |
| Government | An organization operated mainly by the government of one or multiple countries / territories. | Any designation that is used by a government such as a ministry, department, secretariat, office, directorate or federal agency. | A government agency for healthcare is categorised as Government. |
| Other | Used in cases where none of the other aforementioned types are suitable. | Anything not belonging to any of the other categories, or where the primary function is unclear. | A university press would fall into this category. |
Geoinformation
Geograpical and location-related information is obtained from the organization's official webpage or other trusted source.
GeoNames
All organizations are manually assigned a GeoNames ID by our curators. From this we derive the city, state (only for Canada, United States and Australia) and country.
Geographical coordinates (in decimal format, e.g. 12.919849, 77.584146) are retrieved from Google Maps and added manually to the respective longitude and latitude fields in Dimensions.
NUTS codes
We are using Eurostat's NUTS classification (Nomenclature of territorial units for statistics), a hierarchical system for dividing up the economic territory of the European Union (EU) and the United Kingdom (UK). The NUTS version currently in use in Dimensions is NUTS 2013, with all 3 levels, providing regional codes and regions' area geographical coordinates.
All the GeoName IDs that have geo coordinates within a NUTS area's coordinates are linked to that specific area.
When displaying the NUTS codes information on an organization page in Dimensions or in the organizations release JSON export, we rely on the organization's city’s GeoName ID, not on the organization’s coordinates.
Special cases
Countries as US states
| Country | State name | State code | New Country |
|---|---|---|---|
| Puerto Rico | Puerto Rico | US-PR | US |
| Guam | Guam | US-GU | US |
| American Samoa | American Samoa | US-AS | US |
| Northern Mariana Islands | Northern Mariana Islands | US-MP | US |
| Minor Outlying Islands | Minor Outlying Islands | US-UM | US |
| American Virgin islands | American Virgin islands | US-VI | US |
Countries as China subdivisions
| Country | State name | State code | New Country |
|---|---|---|---|
| Hong Kong (Xiangang) | Hong Kong | CN-91 | CN |
| Macau | Macau | CN-92 | CN |
External IDs and links
For each organization record we add and maintain the following external IDs and links:
Institute links
When an official website exits for the organization the link is added.
Added when an organisation is created and updated during reviews.
Wikipedia
When a page exists for the organization in the English language version of Wikipedia the link is added.
Added when an organization is created and updated during reviews.
Dimensions
Automatically created link to to the research organization filter within the Publications content type in Dimensions.
ISNI
The International Standard Name Identifier (ISNI) is a 16-digit identifier added to a record where available.
Added when an organisation is created and updated during reviews.
Open Funder Registry (formerly FundRef)
The Crossref Open Funder Registry is an open registry providing identifiers for grant-giving organizations.
Added when an organisation is created and updated during reviews.
ROR
The Research Organization Registry (ROR) was initially seeded with the Dimensions organizational data. Thus, most ROR records already have GRID IDs from Dimensions and vice versa.
How are new ROR IDs added to Dimensions organization records?
As part of every new organization data release we identify new ROR records, and examine each one of these for relevance to Dimensions, according to our inclusion criteria. Records that meet these criteria are added to Dimensions along with the relevant metadata.
How is the ROR metadata maintained in Dimensions?
We are also regularly checking for changes to ROR records which are already mapped to Dimensions organization records. When a mapped record changes in ROR, we examine if that change is also relevant for Dimensions and update relevant metadata as applicable.