Schema Scope
Citation Topics are algorithmically derived citation clusters (using an algorithm developed by CWTS, Leiden). They form a three-level document-level hierarchical classification system:
Macro Citation Topics – broad scientific domains
Meso Citation Topics – more specific thematic areas within each macro domain
Micro Citation Topics – fine-grained research areas at the level of closely related publications
According to the 2026 clustering, the schema covers:
- Macro topics: 10
- Meso topics: 326
- Micro topics: 2,493
The full 2026 hierarchy and its contents are available for download:
Download the 2026 schema (with descriptions)
2026 update
As part of the April 2026 dataset update, a new version of the Citation Topics clustering and labelling has been released.
Macro- and Meso-topics
There are no changes to the composition or labels of the Macro Citation Topics compared with the previous release.
At the Meso Citation Topic level, the labels of 37 meso-topics were changed.
Micro-topics: annual re-clustering and label review
In 2025, generative AI was introduced to support the assignment of labels to Micro Citation Topics. The same process was repeated as part of the 2026 update.
Generative AI was used to analyse the publication titles belonging to each micro citation topic and generate a candidate label that:
- accurately represents the full domain covered by the micro-topic — that is, the complete set of clustered publications; and
- remains a logical subset of the corresponding meso-topic.
The existing editorial labels were then systematically compared with the newly generated candidate labels, micro-topic by micro-topic. For each micro-topic, the label, either the existing editorial label or the newly generated label, that best described the scientific content of the cluster and fitted consistently within the macro/meso hierarchy was selected.
As a result of this process:
- 2080 micro citation topics retained their existing labels.
- 370 micro citation topic labels were changed
- 43 new micro citation topics were added
- 28 micro citation topics were deleted
- 3 micro citation topics were associated with new macro- and meso-citation topics.
Micro-topic descriptions
To improve interpretability, descriptions were introduced at the micro-topic level in the 2025 schema. These descriptions continue to be provided with every annual Citation Topics update.
The descriptions are not available within the product itself. They are included only in the downloadable Citation Topics 2026 schema file: Citation Topics 2026 update.
For each Micro Citation Topic, the schema provides:
-
Micro label
A short label (typically 1–3 words) naming the micro-topic. -
Longer label
A more descriptive label (up to ~8 words) that adds context to the short micro label. -
Description
A richer textual description (up to ~150 words) explaining the content and scope of the micro-topic.
On average, a Micro Citation Topic clusters approximately 26,000 publications. The description clarifies the types of research, methods and subjects typically included in the cluster. It is intended to help non-experts understand the topic and support better interpretation in analyses, evaluations and communications.
These descriptions are included in the 2026 schema file and are intended to make the micro-level classification more transparent and easier to use.
Schema Type
Named document clusters based on cited and citing relationships between documents.
Method
All documents from 1980-present were algorithmically clustered where possible, based on their cited and citing relationships using a Leiden-type community detection algorithm. The algorithm includes parameters to enforce clustering and maintain minimum cluster size; cited relationships to pre-1980 documents are included. The methodology and algorithmic details are described in:
Documents are only assigned to a single topic.; all document types are considered. Not all documents will have been successfully assigned to a topic, and by default these are not displayed (you can display them using the checkbox in the Research Area filter). Approximately 75% of all documents from 1980 onward are assigned to a topic. This is >90% for article and review document types.
Macro- and meso-topics are manually labeled based on their contents. Micro-topics were algorithmically labeled with their most significant keyword. As Citation Topics are based on citation relationships and not the content or subject matter of their constituent documents, topics are labeled by inference and any name may not be descriptive for every document in the topic.
Each topic is also labeled with a permanent numerical prefix to identify the precise topic. For example, the micro-topic 2.114.321 OLEDs is a sub-topic of the meso topic 2.114 Organic Semiconductors, which is a sub-topic of the macro topic 2 Chemistry.
New documents are added to the existing topics based on their cited references, with each monthly data update. Each year, we perform a full clustering update, which will retain the Citation Topic structure but individual documents may move between micro topics and entirely new micro topics may be added.
Citation Topics
Macro Topics
- Clinical & Life Sciences
- Chemistry
- Agriculture, Environment & Ecology
- Electrical Engineering, Electronics & Computer Science
- Physics
- Social Sciences
- Engineering & Materials Science
- Earth Sciences
- Mathematics
- Arts & Humanities