IDCite: A Large-Scale Multidisciplinary Citation Intent Dataset for Scholarly Knowledge Discovery

IDCite IDCite is a large-scale multidisciplinary citation dataset designed to support research on citation contexts, citation intent, scholarly information retrieval, citation recommendation, scientometrics, and scholarly knowledge graphs. Unlike conventional citation datasets that primarily represent citations as links between papers, IDCite is organized around citation events connecting citing papers to highly cited seed papers while preserving semantic and bibliographic evidence. Each citation event can incorporate citation contexts, automatically generated citation-intent annotations, publication metadata, disciplinary information, and normalized scholarly entities. The current Version 4 release contains: - 1,857,503 citation events- 1,467,045 citing papers- 23,479 highly cited seed papers- 21 Essential Science Indicators (ESI) fields- 31 observed citation-intent labels- 3,420,132 scholarly knowledge graph nodes- 6,851,623 scholarly knowledge graph edges The resource additionally provides normalized representations of authors, affiliations, journals, scientific fields, cities, countries, and citation intents. Version 4 is the version described in the accompanying article and supersedes all earlier versions. Users should use Version 4. What Changed in Version 4 Version 4 removes all Scopus-derived content from the release. All released bibliographic metadata now originates from openly licensed sources. - Seed-paper metadata (titles, publication dates, venues, first authors, affiliations, geographic information, citation counts, bibliographic details, open-access status, and funders), previously obtained from Scopus, is now retrieved from OpenAlex (CC0) by DOI.- Scopus-specific identifiers and fields (e.g., Scopus EIDs, author and affiliation identifiers, funding fields) have been removed.- The raw Scopus and journal-level sampling archives distributed with earlier versions are no longer included.- The normalized entity tables (authors, affiliations, affiliation_geo, cities, countries, journals) and the knowledge graph were rebuilt from the OpenAlex-based seed metadata, using the same construction logic as Version 3.- OpenAlex identifiers (works, sources, authors, institutions) and ROR identifiers were added to the seed-paper tables to improve interoperability. Unchanged from Version 3: the seed-paper set (23,479 DOIs), all citation events, citation contexts, citing-paper records, citation-intent annotations, field and intent tables, and the human-evaluation archive. Relationship to Previous Releases IDCite represents the fourth stage in the evolution of the scholarly citation resources distributed through this Zenodo record. Version 1 — MDCite MDCite was released as the original large-scale multidisciplinary citation-context dataset. It focused primarily on citation contexts and citation-intent information for citation-aware scholarly analysis. Version 2 — MDContextCite and EdgeCite MDContextCite extended the initial resource into a broader scholarly citation database and dissemination framework. Within MDContextCite: - MDCite served as the core citation-context dataset.- EdgeCite provided an additional retrieval-oriented variant designed for leakage-controlled large-scale citation retrieval experiments. Version 3 — IDCite IDCite reorganized and extended the underlying scholarly citation resources around citation events, normalized scholarly entities, and an ontology-ready scholarly knowledge graph representation. IDCite should not be interpreted as a simple renaming of MDContextCite. Version 4 — IDCite (current) Version 4 is the current release of IDCite. It preserves the Version 3 data model and citation content while replacing all Scopus-derived metadata with openly licensed OpenAlex metadata, so that every released component can be redistributed under the terms stated below. Earlier versions are retained in the Zenodo version history for provenance only and are superseded by Version 4. Data Sources IDCite integrates information from multiple scholarly infrastructures. The table below summarizes how each source was used and whether its content is redistributed. 🔹 OpenAlex (redistributed; CC0 1.0) OpenAlex provides citation links between citing papers and seed papers, and all released seed-paper bibliographic metadata, including titles, publication dates, venues, first authors, affiliations, institution locations, citation counts, bibliographic details, open-access status, and funders. 🔹 Semantic Scholar (redistributed; ODC-BY 1.0) The Semantic Scholar Academic Graph API provides citation contexts and citing-paper metadata (titles, years, venues, and open-access PDF links). 🔹 Automatic Citation Intent Classification (authors' contribution; CC BY 4.0) Citation-intent annotations are generated automatically at scale using the SynIntent classifier. The resulting annotations should be interpreted as model-derived semantic annotations / weak semantic supervision, rather than manually curated gold-standard labels for every citation event. 🔹 Scopus (used for selection only; not redistributed) Scopus citation counts were used, under the authors' institutional subscription, solely to rank publications within each journal and select the top 5% most-cited papers as seed papers. No Scopus records, identifiers, or citation counts are included in this release; seed papers are identified by DOI only. 🔹 Web of Science / Journal Citation Reports (used for selection only; not redistributed) Web of Science subject categories and the 2024 Journal Citation Reports were used for multidisciplinary journal-stratified sampling. One representative Web of Science category is associated with each of the 21 ESI fields, and five representative Q1 journals are selected within each category. No JCR or Web of Science metrics (e.g., impact factors or quartile values) are included in the release; only journal names and subject-category labels are used. Dataset Construction Pipeline The IDCite construction workflow consists of the following major stages: 1. Journal-stratified sampling across 21 ESI fields and 21 representative Web of Science categories.2. Selection of five representative Q1 journals per category, resulting in 105 journals.3. Independent selection of the top 5% most-cited papers within each journal (ranked by Scopus citation counts) to construct the seed-paper corpus.4. Retrieval of seed-paper bibliographic metadata from OpenAlex by DOI.5. Citation-link harvesting and citation-event construction between citing papers and seed papers.6. Citation-context collection and automatic citation-intent annotation.7. Normalization of publications, journals, authors, affiliations, geographic entities, scientific fields, and citation intents.8. Conversion of normalized entities and relationships into an ontology-ready scholarly knowledge graph. The journal-stratified sampling strategy is intended to preserve disciplinary heterogeneity and reduce the dominance of citation-intensive disciplines that would arise from applying a single global citation threshold. Because the seed corpus focuses on highly cited publications, IDCite primarily characterizes citation behavior around influential scholarly works and should not automatically be interpreted as representative of citation behavior surrounding moderately or rarely cited publications. Citation Intent Representation IDCite contains 31 observed citation-intent labels, consisting of 30 valid observed categories and one missing label retained for reproducibility. These labels include seven canonical citation intents: - background- uses- similarities- differences- motivation- extends- future work Composite labels are also retained in the released data. Composite intent labels can be decomposed into their constituent canonical intents, allowing a citation event to contribute to more than one semantic intent category. Because citation-intent labels are automatically inferred, annotation quality is model-dependent and may vary across disciplines and citation contexts. Human Evaluation of Citation-Intent Annotations For the reliability of the automatically generated citation-intent annotations, the release includes a human-evaluation archive: - `human_evaluation_Tuan_Anh_Phan.zip` This archive contains human evaluation results for citation-intent predictions on a representative multidisciplinary sample of the corpus. The evaluation is organized across 21 disciplinary categories and the 7 canonical citation intents: `background`, `uses`, `similarities`, `differences`, `motivation`, `extends`, and `future_work`. For each discipline and citation-intent category, a separate CSV file is provided. Each CSV contains evaluated citation contexts and the following three fields: - `citation context` — the citation-context text evaluated- `Prediction from classification model` — the citation intent predicted by the automatic classification model- `Prediction from human` — the corresponding citation intent assigned through human evaluation In total, the archive contains 147 CSV evaluation files (21 disciplines × 7 canonical intents), providing direct model–human comparison data across disciplinary and citation-intent categories. The human-evaluation files complement the benchmark-based evaluation of the citation-intent classifier and provide additional evidence for assessing the quality of the automatically inferred semantic annotations released with IDCite. IDCite Dataset Files The Version 4 release provides 17 Parquet files covering citation events, publications, normalized scholarly entities, and graph representations. Citation Event Files - `citation_events.parquet`- `citation_events_enriched.parquet`- `citation_events_normalized.parquet` These files contain 1,857,503 citation events at different levels of enrichment and normalization. In `citation_events_enriched.parquet`, the `cited_*` columns carry OpenAlex-based seed-paper metadata. Citing Paper Files - `citing_papers.parquet`- `citing_papers_normalized.parquet` These files contain metadata for 1,467,045 citing papers. Seed Paper Files - `seed_cited_papers.parquet`- `seed_cited_papers_normalized.parquet` These files contain metadata for 23,479 highly cited seed papers. Bibliographic metadata were retrieved from OpenAlex for 23,252 seed papers (99.03%). The remaining 227 seed papers are not indexed in OpenAlex and are retained with their DOI and sampling labels (ESI field, Web of Science category, and journal) only. Seed-paper citation counts are OpenAlex counts at the time of metadata retrieval and may differ from the Scopus counts used for seed selection. Normalized Scholarly Entity Files - `authors.parquet` (19,865 first-author entities of seed papers)- `affiliations.parquet` (4,074 entities, first listed institution of each seed paper)- `affiliation_geo.parquet` (4,088 affiliation–city–country mappings)- `journals.parquet` (46,234)- `fields.parquet` (21)- `intents.parquet` (31)- `cities.parquet` (1,760)- `countries.parquet` (120) These lookup tables provide dataset-level normalized identifiers for scholarly entities and support consistent joins across the IDCite files. Scholarly Knowledge Graph Files - `kg_nodes.parquet`- `kg_edges.parquet` The graph representation contains: - 3,420,132 nodes- 6,851,623 typed edges Scholarly Knowledge Graph IDCite provides a heterogeneous scholarly knowledge graph derived from the normalized citation-event infrastructure. The graph represents entities including: - citation events- citing and cited publications- authors- affiliations- journals- scientific fields- cities- countries- citation intents Typed relations encode citation-event linkage, authorship, affiliation, publication venue, geographic association, disciplinary assignment, and citation-intent information. The knowledge graph is provided as a supplementary representation of IDCite and does not replace the tabular citation-event files. Ontology-ready Representation In IDCite, ontology-ready means that major scholarly entities are represented using normalized dataset-level identifiers and connected through explicit typed relationships. The term does not imply that a single external formal ontology has already been imposed on every entity. Instead, IDCite provides a structured graph-ready representation that can be mapped to or extended with formal scholarly ontologies. Intended Research Uses IDCite is designed to support research including: - citation context analysis- citation intent analysis and classification- scholarly information retrieval- citation recommendation- scientometric and bibliometric analysis- cross-disciplinary citation analysis- scholarly knowledge graph construction and analysis- graph representation learning- link prediction- research trend analysis- scholarly knowledge discovery The combination of contextual semantic evidence and structured scholarly relationships enables analyses that cannot be performed using paper-to-paper citation links alone. CitationHub CitationHub is the interactive dashboard and visualization platform built on top of IDCite. CitationHub is not a separate Zenodo dataset version. IDCite provides the structured data layer, while CitationHub provides the user-facing exploration and visualization layer. CitationHub supports interactive exploration of: - citation relationships- citation contexts- citation intents- disciplinary patterns- scholarly entities- knowledge graph relationships CitationHub search interface URL: https://citation-hub-website.vercel.app Reproducibility The released Parquet files, normalized lookup tables, documentation, and graph files are designed to support direct reuse without requiring reconstruction of the entire upstream pipeline. Full reconstruction may require: - institutional access to Scopus (only to reproduce the seed-paper ranking)- access to Web of Science / Journal Citation Reports (only to reproduce journal selection)- access to OpenAlex- access to Semantic Scholar API credentials and proprietary database access are not distributed with IDCite. The construction code, including the Version 4 OpenAlex metadata step, is available at https://github.com/kecau/MDCite. Because upstream scholarly infrastructures are continuously updated, IDCite should be interpreted as a snapshot. Citation links, citation contexts, and citing-paper records correspond to the November 2025 collection period; seed-paper metadata were retrieved from OpenAlex in November 2026. Licensing and Responsible Use IDCite is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This license covers the authors' original contributions: the citation-event structure, normalized entity identifiers, automatically generated citation-intent annotations, the knowledge-graph representation, and the human-evaluation archive. Information obtained from third-party sources remains subject to the terms of those sources: - OpenAlex metadata (citation links and seed-paper bibliographic metadata) are released by OpenAlex under CC0 1.0 and may be reused without restriction.- Semantic Scholar data (citation contexts and citing-paper metadata) are made available under ODC-BY 1.0. Reuse requires attribution to Semantic Scholar and citation of Kinney et al. (2023), "The Semantic Scholar Open Data Platform." Citation contexts are short excerpts of the citing publications; copyright in the underlying text remains with the respective authors and publishers, and the CC BY 4.0 license of IDCite does not extend to these excerpts, which are provided for research purposes.- Scopus and Web of Science / Journal Citation Reports were used only for seed-paper and journal selection. No content from these services is redistributed. "Citation-based indicators derived from IDCite should not be used as the sole basis for evaluating individual researchers, institutions, journals, scientific fields, or countries" Citation If you find this work useful and are willing to use IDCite, please cite: Nam, Seohyun. (2026).IDCite: A Large-Scale Multidisciplinary Citation Intent Dataset for Scholarly Knowledge Discovery (Version 4).Zenodo.https://doi.org/10.5281/zenodo.23914564 BibTeX: @dataset{idcite2026, title = {IDCite: A Large-Scale Multidisciplinary Citation Intent Dataset for Scholarly Knowledge Discovery}, author = {Nam, Seohyun}, year = {2026}, version = {4}, publisher = {Zenodo}, doi = {10.5281/zenodo.23914564} } Version Information - Concept DOI: 10.5281/zenodo.18410049- Version 1 — MDCite: 10.5281/zenodo.18410050- Version 2 — MDContextCite / EdgeCite: 10.5281/zenodo.18536895- Version 3 — IDCite: 10.5281/zenodo.20796923- Version 4 — IDCite (current): 10.5281/zenodo.23914564 Version 4 removes all Scopus-derived content and replaces it with openly licensed OpenAlex metadata, while preserving the Version 3 citation events, citation contexts, citation-intent annotations. Earlier versions remain accessible through the Zenodo version history for provenance only and are superseded by Version 4.