eduKateSG · Why Science?
A citation count becomes useful when every edge can be traced
Read OpenCitations through citing and cited entities, OMIDs, Open Citation Identifiers, source collections, provenance records, APIs and downloadable snapshots.
References let scientific ideas travel across time. A paper can build on an earlier method, compare a result, reuse data or dispute a conclusion. OpenCitations publishes open bibliographic and citation data so those directed connections can be queried, downloaded and examined.
The current OpenCitations service separates OpenCitations Meta, which contains bibliographic metadata for works involved in citations, from OpenCitations Index, which stores citation links gathered from several sources. OpenCitations Meta Identifiers, or OMIDs, give entities internal identities across source identifiers.
Did you know? A citation itself can have an Open Citation Identifier, or OCI. That means the relationship from one work to another can be addressed as an entity rather than appearing only as an anonymous number in a total.
This guide treats citation data as a discovery and provenance resource. A visible citation says that one indexed work references another under the available data; it does not say the citing authors agreed, that the citation was positive, or that either conclusion is correct.
Chapter index
Chapters 1–9: citation entities, identifiers and direction
- 1. A citation has direction
- 2. Bibliographic entities need identities
- 3. The OCI identifies the link
- 4. Meta and Index answer different questions
- 5. Source collections preserve origin
- 6. Provenance is published separately
- 7. DOIs are useful but not universal
- 8. References and citations are opposite views
- 9. Counts inherit coverage limits
Chapters 10–17: sources, APIs, snapshots and limits
A citation is a directed link from a citing work to a cited work. The citing entity contains a reference to the cited entity. Reversing that direction changes the scholarly history and any derived count. This turns reference checking into a transparent science skill.
Try this
Evidence question: Which work is citing and which is cited? 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 the two identifiers with a labelled arrow. 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 citation has direction, its identifier or version context, and one limit that must remain visible.
Open citation data becomes useful when identifiers and provenance remain attached. OpenCitations Meta assigns OMIDs to bibliographic resources and responsible agents while retaining external identifiers such as DOI where present. It also keeps a graph edge from being mistaken for a complete reading of either paper.
Try this
Evidence question: Which OMID and external ID refer to this work? 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: Resolve both routes and keep their identifier schemes explicit. 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 bibliographic entities need identities, its identifier or version context, and one limit that must remain visible.
Counts summarise available links; they do not measure truth or quality by themselves. An Open Citation Identifier denotes a particular citation relationship. It is not the identifier for either endpoint work. Students can explore influence while respecting database coverage and version limits.
Try this
Evidence question: Does this value identify an edge or a document? 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: Store citation, citing work and cited work in separate fields. 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 the oci identifies the link, its identifier or version context, and one limit that must remain visible.
A reproducible query records the endpoint, identifier form, filters and retrieval date. OpenCitations Meta describes works and agents; OpenCitations Index describes citations among bibliographic entities. A complete investigation may need both. The route stays useful because another reader can repeat it and inspect both works.
Try this
Evidence question: Which dataset supplied this field? 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: Label every retrieved field as Meta or Index. 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 meta and index answer different questions, its identifier or version context, and one limit that must remain visible.
A citation is a directed link from a citing work to a cited work. The unified index combines citation data from collections such as COCI, DOCI and other sources. Collection identity helps a reader understand coverage and update routes. This turns reference checking into a transparent science skill.
Try this
Evidence question: Which collection contributed this citation? 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 collection name before combining rows. 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 source collections preserve origin, its identifier or version context, and one limit that must remain visible.
Open citation data becomes useful when identifiers and provenance remain attached. OpenCitations downloads include provenance data that describes how citation and metadata statements were created or sourced. Provenance supports audit but does not replace the original papers. It also keeps a graph edge from being mistaken for a complete reading of either paper.
Try this
Evidence question: Which provenance record supports the edge? 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: Pair one citation row with its provenance row. 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 provenance is published separately, its identifier or version context, and one limit that must remain visible.
Counts summarise available links; they do not measure truth or quality by themselves. Many queries use DOI identifiers, while OpenCitations also supports other identifiers and OMIDs. Works without a DOI should not be assumed absent from scholarship. Students can explore influence while respecting database coverage and version limits.
Try this
Evidence question: Which identifier forms does the chosen endpoint accept? 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: Try the canonical supported prefix instead of a title guess. 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 dois are useful but not universal, its identifier or version context, and one limit that must remain visible.
A reproducible query records the endpoint, identifier form, filters and retrieval date. A references request asks what one work cites; a citations request asks which indexed works cite it. The same graph edge appears from opposite traversal directions. The route stays useful because another reader can repeat it and inspect both works.
Try this
Evidence question: Which traversal answers the research question? 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: Retrieve both views for a sample DOI and name the difference. 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 references and citations are opposite views, its identifier or version context, and one limit that must remain visible.
A citation is a directed link from a citing work to a cited work. Citation-count and reference-count endpoints summarise links in the current index. A total depends on contributing sources, matching and snapshot date. This turns reference checking into a transparent science skill.
Try this
Evidence question: What exactly is counted, in which release? 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: Attach service, endpoint and date to every count. 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 counts inherit coverage limits, its identifier or version context, and one limit that must remain visible.
Open citation data becomes useful when identifiers and provenance remain attached. Citation metadata can include a timespan between publication dates. A negative or missing value may reflect dates or version issues needing investigation. It also keeps a graph edge from being mistaken for a complete reading of either paper.
Try this
Evidence question: Which dates produced this interval? 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 both publication dates before interpreting the span. 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 timespan is a derived property, its identifier or version context, and one limit that must remain visible.
Counts summarise available links; they do not measure truth or quality by themselves. Self-citation indicators depend on shared authorship or venue definitions encoded by the service. They are analytic labels rather than ethical verdicts. Students can explore influence while respecting database coverage and version limits.
Try this
Evidence question: What matching rule produced this flag? 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: Read the field definition before reporting a self-citation rate. 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 self-citation labels need definitions, its identifier or version context, and one limit that must remain visible.
A reproducible query records the endpoint, identifier form, filters and retrieval date. The Index V2 API accepts prefixed identifiers and exposes citation, count, citations and references operations. Legacy endpoints can differ and may be phased out. The route stays useful because another reader can repeat it and inspect both works.
Try this
Evidence question: Which API version generated the 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: Save the base URL and version with the result. 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 api versions matter, its identifier or version context, and one limit that must remain visible.
A citation is a directed link from a citing work to a cited work. Require, filter, sort and format parameters can exclude rows or change ordering. A filtered result should not be described as the whole index. This turns reference checking into a transparent science skill.
Try this
Evidence question: Which parameter removed or reordered records? 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: Preserve the complete request URL. 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 filters reshape the returned table, its identifier or version context, and one limit that must remain visible.
CHAPTER 14 OF 17 · Interpret counts carefully
14. Rate limits affect retrieval design
Open citation data becomes useful when identifiers and provenance remain attached. The current Index documentation states a per-IP request limit and recommends dumps for large-scale retrieval. Responsible clients avoid repeated calls that burden shared infrastructure. It also keeps a graph edge from being mistaken for a complete reading of either paper.
Try this
Evidence question: Is this task an API lookup or bulk-analysis job? 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: Choose the dump when the planned request volume is large. 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 rate limits affect retrieval design, its identifier or version context, and one limit that must remain visible.
Counts summarise available links; they do not measure truth or quality by themselves. Downloadable dumps provide citation, bibliographic and provenance data for a dated release in several formats. Later releases can add sources or corrections. Students can explore influence while respecting database coverage and version limits.
Try this
Evidence question: Which snapshot and format were analysed? 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 release date and checksum or filename in the method. 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 snapshots freeze a release, its identifier or version context, and one limit that must remain visible.
CHAPTER 16 OF 17 · Interpret counts carefully
16. Open data still needs field literacy
A reproducible query records the endpoint, identifier form, filters and retrieval date. CSV, RDF, JSON-LD and Scholix representations express related information with different structures. Conversion can lose context if identifiers or provenance are dropped. The route stays useful because another reader can repeat it and inspect both works.
Try this
Evidence question: Which fields must survive transformation? 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: Round-trip a small sample and compare identifiers and direction. 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 open data still needs field literacy, its identifier or version context, and one limit that must remain visible.
CHAPTER 17 OF 17 · Interpret counts carefully
17. Citation infrastructure supports careers
A citation is a directed link from a citing work to a cited work. Bibliometricians, librarians, data engineers, research evaluators and developers build and audit citation graphs. Their shared responsibility is transparent coverage and interpretation. This turns reference checking into a transparent science skill.
Try this
Evidence question: Which decision could be harmed by treating a count as quality? 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 a citation receipt for a student 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 citation infrastructure supports careers, its identifier or version context, and one limit that must remain visible.
Contents · Previous chapter · Continue at the Science Learning Hub
Count only after you can trace the link
OpenCitations makes scholarly references available as open, queryable graph data. Its strength lies in addressable citation entities, reusable identifiers, published provenance and dated releases.
Keep the OCI or endpoint pair, OMIDs, external identifiers, source collection, API version, filters and snapshot date together. Then read the citing context and both underlying works before explaining what the citation means.
| Reading object | What it can support | Question to keep visible |
|---|---|---|
| Citation edge | Directed reference between works | Which endpoint is citing? |
| OMID | OpenCitations entity identity | Which external identifier maps to it? |
| Provenance | Origin and processing context | Which source collection supplied the link? |
| Count | Summary of indexed edges | Which release and coverage limits apply? |
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
- OpenCitations home — accessed 25 October 2026.
- OpenCitations API gateway — accessed 25 October 2026.
- OpenCitations Index V2 API — accessed 25 October 2026.
- OpenCitations data downloads — accessed 25 October 2026.
- OpenCitations publications and infrastructure description — accessed 25 October 2026.