eduKateSG · Why Science?
A research graph becomes useful when every connection keeps its identity
Read OpenAlex works through authorships, sources, institutions, topics, external identifiers, corpus choices and repeatable API queries without confusing discovery metadata with scientific proof.
Scientific literature is spread across publishers, repositories, preprint servers and institutional systems. OpenAlex provides a connected index of scholarly entities so a learner can move among works, authors, publication sources, institutions, publishers, funders and topics without starting from a title search every time.
The current OpenAlex data reference distinguishes native entities, vocabulary entities, component entities and user-created entities. Works, authors, sources and institutions receive OpenAlex identifiers; an authorship is a structured component inside a work. That distinction matters because not every visible object has the same identity or curation process.
Did you know? OpenAlex maps its entities to widely used external identifiers where available: DOI for works, ORCID for authors, ISSN-L for sources and ROR for institutions. These mappings create useful bridges, but each identifier still belongs to a different registry and object type.
This guide treats OpenAlex as a discovery and analysis graph. Its records can help formulate questions and retrieve candidate evidence. The publication, dataset, methods, correction notice and source repository still determine what a scientific claim can support.
Chapter index
Chapters 1–9: entities, identifiers and authorships
- 1. A work is a scholarly object
- 2. Native IDs encode entity type
- 3. External identifiers create bridges
- 4. Authorships join people and institutions
- 5. Author records are disambiguated
- 6. Sources are publication containers
- 7. Institutions need organisation identity
- 8. Locations describe availability
- 9. Topics describe aboutness
Chapters 10–17: queries, counts, provenance and limits
- 10. Citation links depend on indexed records
- 11. Merged IDs preserve old routes
- 12. Full and dehydrated objects differ
- 13. The corpus choice changes counts
- 14. Search and filter answer different questions
- 15. Paging is part of the method
- 16. Selected fields change the response
- 17. Graph literacy supports science careers
A scholarly graph is useful when every node keeps its identity and source. The Works entity covers scholarly outputs represented in the OpenAlex corpus and connects them with identifiers, authorships, locations, topics and citations. A work record is metadata about an output rather than the complete output. That distinction turns discovery into a reproducible science skill.
Try this
Evidence question: Which exact work and output type does this record represent? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Save the OpenAlex work ID, DOI when present, title and type before reading its counts. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for a work is a scholarly object, its identifier or version context, and one limit that must remain visible.
One search result can combine a work, people, institutions and a publication venue. Short native IDs use prefixes such as W for works, A for authors, S for sources and I for institutions. The fully qualified OpenAlex URL is the clearest reusable form. It also keeps a convenient index from becoming an unsupported authority claim.
Try this
Evidence question: Does the copied identifier still reveal its entity type and resolve correctly? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Convert one short ID to a full OpenAlex URL and test it. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for native ids encode entity type, its identifier or version context, and one limit that must remain visible.
Identifiers let a reader move between systems without relying on ambiguous names. An entity may carry an ids object linking to DOI, ORCID, ROR, ISSN or other established identifiers. Missing mappings should be recorded as absent rather than guessed. A learner can follow the graph while preserving uncertainty and provenance.
Try this
Evidence question: Which registry owns each external identifier? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Make a two-column map from OpenAlex ID to verified external ID. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for external identifiers create bridges, its identifier or version context, and one limit that must remain visible.
Counts and rankings become meaningful only after the query and corpus boundary are recorded. A work embeds authorship objects that connect author identities with raw affiliation strings and matched institutions. This is a structured assertion, not automatic proof of every employment detail. The result is a research route that another person can inspect and repeat.
Try this
Evidence question: Which author, institution and raw affiliation string are linked? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Inspect one authorship and keep the raw and matched forms side by side. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for authorships join people and institutions, its identifier or version context, and one limit that must remain visible.
A scholarly graph is useful when every node keeps its identity and source. OpenAlex constructs author entities from publication evidence and other identifiers. Name similarity alone can merge or split people incorrectly, so ORCID and work history deserve checking. That distinction turns discovery into a reproducible science skill.
Try this
Evidence question: What evidence distinguishes this author from a namesake? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Compare the OpenAlex author page with an authenticated ORCID record. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for author records are disambiguated, its identifier or version context, and one limit that must remain visible.
One search result can combine a work, people, institutions and a publication venue. A source can represent a journal, repository, conference series or another place where works appear. Source type and identifiers such as ISSN-L help distinguish similarly named venues. It also keeps a convenient index from becoming an unsupported authority claim.
Try this
Evidence question: What kind of source is this and which identifier anchors it? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Record source type, host organisation if shown and canonical external ID. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for sources are publication containers, its identifier or version context, and one limit that must remain visible.
Identifiers let a reader move between systems without relying on ambiguous names. Institution entities use names, locations, relationships and external mappings such as ROR. An institution match should not be inferred from an acronym alone. A learner can follow the graph while preserving uncertainty and provenance.
Try this
Evidence question: Does the ROR mapping resolve to the intended organisation? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Verify one institution with its ROR record and location. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for institutions need organisation identity, its identifier or version context, and one limit that must remain visible.
Counts and rankings become meaningful only after the query and corpus boundary are recorded. A work can have one or more locations representing where versions or manifestations are hosted. A location can carry source, landing-page and open-access information. The result is a research route that another person can inspect and repeat.
Try this
Evidence question: Which location provides the object you actually inspected? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Record the exact landing page and version rather than citing the work graph generally. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for locations describe availability, its identifier or version context, and one limit that must remain visible.
A scholarly graph is useful when every node keeps its identity and source. Topics and higher-level fields classify what works are about. Classification supports filtering and exploration but does not replace reading the title, abstract, methods or full text. That distinction turns discovery into a reproducible science skill.
Try this
Evidence question: Which topic label is useful, and which nuance does it omit? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Compare the assigned topic with the paper’s own research question. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for topics describe aboutness, its identifier or version context, and one limit that must remain visible.
One search result can combine a work, people, institutions and a publication venue. Referenced-works and cited-by connections describe links available in the OpenAlex graph. Coverage and matching affect counts, so citation totals are not universal measures of quality. It also keeps a convenient index from becoming an unsupported authority claim.
Try this
Evidence question: What corpus and retrieval date produced the count? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Write a narrow count statement with the database and date attached. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for citation links depend on indexed records, its identifier or version context, and one limit that must remain visible.
Identifiers let a reader move between systems without relying on ambiguous names. When OpenAlex determines that two native IDs describe the same entity, one can redirect to the canonical entity. The redirect maintains a route while signalling that identity decisions can change. A learner can follow the graph while preserving uncertainty and provenance.
Try this
Evidence question: Did the requested ID redirect to another ID? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Record both requested and canonical IDs in a reconciliation note. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for merged ids preserve old routes, its identifier or version context, and one limit that must remain visible.
Counts and rankings become meaningful only after the query and corpus boundary are recorded. Embedded entity references are often dehydrated stubs with a small set of fields. Fetching the entity endpoint returns the full object needed for deeper checks. The result is a research route that another person can inspect and repeat.
Try this
Evidence question: Is this a compact reference or the complete entity response? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Follow one embedded author ID to its full record before using additional fields. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for full and dehydrated objects differ, its identifier or version context, and one limit that must remain visible.
A scholarly graph is useful when every node keeps its identity and source. The API documentation distinguishes the curated core corpus from a larger opt-in expansion corpus. A query without its corpus setting cannot be reproduced precisely. That distinction turns discovery into a reproducible science skill.
Try this
Evidence question: Did the request use the default core corpus or corpus=all? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Run the same grouped query under both settings and explain the difference. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for the corpus choice changes counts, its identifier or version context, and one limit that must remain visible.
CHAPTER 14 OF 17 · Evaluate the route
14. Search and filter answer different questions
One search result can combine a work, people, institutions and a publication venue. Text search finds records by indexed text, while filters constrain structured fields and entity IDs. Official guidance recommends filtering by IDs rather than ambiguous names when possible. It also keeps a convenient index from becoming an unsupported authority claim.
Try this
Evidence question: Which part of the question belongs in search and which in filters? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Rewrite a name filter using a verified OpenAlex entity ID. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for search and filter answer different questions, its identifier or version context, and one limit that must remain visible.
Identifiers let a reader move between systems without relying on ambiguous names. List endpoints return result windows with metadata such as total count, page and page size. A first page is not the complete result set. A learner can follow the graph while preserving uncertainty and provenance.
Try this
Evidence question: Which window of the query was retrieved? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Save per-page, cursor or page details beside the exported rows. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for paging is part of the method, its identifier or version context, and one limit that must remain visible.
Counts and rankings become meaningful only after the query and corpus boundary are recorded. API clients can request only chosen fields for efficiency. Omitted fields in a selected response are not necessarily absent from the underlying record. The result is a research route that another person can inspect and repeat.
Try this
Evidence question: Was a field missing from the entity or excluded by select? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Repeat one request without field selection before calling a value absent. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for selected fields change the response, its identifier or version context, and one limit that must remain visible.
A scholarly graph is useful when every node keeps its identity and source. Librarians, bibliometricians, research analysts, developers and policy teams use scholarly graphs to connect outputs and organisations. Their work depends on reproducible queries and careful interpretation. That distinction turns discovery into a reproducible science skill.
Try this
Evidence question: Which decision could be distorted by an ambiguous author or corpus boundary? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Design an OpenAlex research receipt for a school literature review. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for graph literacy supports science careers, its identifier or version context, and one limit that must remain visible.
Contents · Previous chapter · Continue at the Science Learning Hub
Use the graph to find evidence, then inspect the evidence itself
OpenAlex can turn a broad science question into a navigable set of works, people, sources and institutions. Its greatest value comes from explicit identifiers and reusable query mechanics.
Keep the entity IDs, external mappings, corpus setting, query, result window and access date together. Then read the underlying object and its current update status before making a scientific claim.
| Reading object | What it can support | Question to keep visible |
|---|---|---|
| Work | Indexed scholarly output | Which underlying object and version were inspected? |
| Authorship | Connection among work, author and affiliation | What raw and matched affiliation evidence is present? |
| Entity ID | OpenAlex identity route | Does an external identifier confirm it? |
| Query | Search, filter and corpus choices | Can another reader reproduce the result? |
Sources and next reading
Substantive resource claims in this article were checked against the primary pages below. A structured record is a route into evidence; it does not by itself prove a scientific conclusion. Recheck the live record, evidence labels and release context before using it in assessed work or research.
For a classroom enquiry, save a short research receipt containing the question, search terms, selected object, exact identifier, version, access date and the evidence boundary you observed. Add the file or page that supports each statement. This small routine turns browsing into accountable science learning and gives a teacher, classmate or future analyst a route they can independently follow.
Before sharing the conclusion, ask another reader to retrace the route without verbal hints. If they reach a different version, file or interpretation, improve the receipt and narrow the claim. Reproducibility begins with this practical check. It teaches careful reading, supports fair collaboration and keeps uncertainty visible while evidence moves from an online service into notes, analysis and a final explanation.
A useful final review separates five questions. What object was found? Which organisation or repository maintains it? What exact evidence was inspected? Which statement does that evidence support? What remains uncertain? Answer each question in one sentence, then connect every answer to a visible field, file or source page. This review catches common mistakes: citing a collection when a file supplied the value, treating the newest release as the analysed release, confusing permission with scientific suitability, or turning a discovery count into a conclusion. It also gives students a calm way to improve work. A missing detail becomes a next research step instead of an invitation to guess. When the route is complete, the final explanation can stay simple because the identifiers, dates, versions, limitations and supporting evidence are already organised underneath it.
Primary sources
- OpenAlex data reference — accessed 25 October 2026.
- OpenAlex works reference — accessed 25 October 2026.
- OpenAlex authors reference — accessed 25 October 2026.
- OpenAlex sources reference — accessed 25 October 2026.
- OpenAlex API reference — accessed 25 October 2026.