SciGalaxy
FeaturesExample mapsPricingTutorial
Log inStart free

NewAI names clusters and writes a downloadable review

Turn a library of papers into a map you can read

Import Web of Science, CNKI, or Scopus, or search OpenAlex. Co-citation, clustering, bursts, and timelines run in your browser. AI writes the review. Your files stay on this computer.

Start a free analysisSee example maps
  • No Java install
  • Chinese databases import directly
  • Export SVG, PDF, and PNG
See example maps#0 medical image segmentation#1 image segmentation#2 computer-aided diagnosis#3 u-net#4 diagnostic#5 limited training#6 feature fusion#7 convolutional neural networks#8 magnetic resonance imaging
Q
0.787
Nodes
422
Clusters
19

Example · medical imaging

Burst 10.1 · 2023–2025

AI review

Example. Keyword co-occurrence on the open OpenAlex sample of deep learning in medical imaging (2015–2025), Pathfinder pruning. Modularity Q = 0.787, 422 nodes, 19 clusters. The map shows the 9 largest: #0 medical image segmentation, #1 image segmentation, #2 computer-aided diagnosis, #3 u-net, #4 diagnostic, #5 limited training, #6 feature fusion, #7 convolutional neural networks, #8 magnetic resonance imaging. Red rings mark Kleinberg bursts.

Q
0.787
Nodes
422
Clusters
19

Example · medical imaging

AI review

Example. Keyword co-occurrence on the open OpenAlex sample of deep learning in medical imaging (2015–2025), Pathfinder pruning. Modularity Q = 0.787, 422 nodes, 19 clusters. The map shows the 9 largest: #0 medical image segmentation, #1 image segmentation, #2 computer-aided diagnosis, #3 u-net, #4 diagnostic, #5 limited training, #6 feature fusion, #7 convolutional neural networks, #8 magnetic resonance imaging. Red rings mark Kleinberg bursts.

Import the export file from these databases

  • Web of Science
  • Scopus
  • CNKI
  • Wanfang
  • CSSCI
  • PubMed
  • OpenAlex

OpenAlex can also be searched online

Features

The bibliometric analyses papers actually use

Networks, clusters, bursts, and time views, in the browser.

  • Cluster map
  • Timeline
  • Timezone
  • Ring
  • Landscape
  • Heatmap
  • t-SNE
  • Geographic map
  • Dual-map overlay
  • Cluster alluvial
#0 medical image segmenta#1 image segmentation#2 computer-aided diagnos

Example · nodes 422 · links 724 · Pathfinder

Co-citation, co-occurrence, collaboration

Document, author, and journal co-citation; keyword co-occurrence and terms; author, institution, and country collaboration; subject categories; document, author, and journal coupling. Twelve node types. Time slicing, g-index, and Pathfinder or MST pruning.

Q 0.787 · 19 clusters · Example

Clusters and automatic names

Louvain clusters, with modularity Q, silhouette S, and LLR / LSI / MI labels. AI can name the clusters.

medical image segmentation
10.12023–2025
long-range dependencies
8.12021–2024
brain tumor classification
6.82022–2023
cnn-transformer hybrid
5.52022–2024
self-attention
5.32021–2024
transformer models
5.12021–2024

Example

Burst detection

Kleinberg bursts find terms and documents that heat up in a span of years. Betweenness, sigma, structural variation (SVA), and entropy are computed too.

201520202025#7#1#2#5#8#0#3#4#6

Timeline and timezone

See when each direction appears, when it peaks, and whether it is still active. Timezone view orders keywords by the year they first show up.

AI review

Example. Keyword co-occurrence on the open OpenAlex sample of deep learning in medical imaging (2015–2025), Pathfinder pruning. Modularity Q = 0.787, 422 nodes, 19 clusters. The map shows the 9 largest: #0 medical image segmentation, #1 image segmentation, #2 computer-aided diagnosis, #3 u-net, #4 diagnostic, #5 limited training, #6 feature fusion, #7 convolutional neural networks, #8 magnetic resonance imaging. Red rings mark Kleinberg bursts.

AI review

AI writes a summary for each cluster and cites only works that are in your data. The review is 7 chapters and downloads as Markdown. AI features send the analysis summary only.

savedrecs.txt · WoS
CNKI.txt · CNKI
scopus.csv · Scopus

scigalaxy.com

Runs in this browser

Files stay on your computer

Parsing and the network run in the browser. The original file is not uploaded. You can also bring in a CiteSpace project folder: data, parameters, aliases, and the exclusion list.

SVGPDFPNGCSVGraphMLGEXFPajekRISAI review

Export for a paper

SVG, PDF, PNG (300 / 600 dpi), CSV, GraphML, GEXF, Pajek, adjacency matrix, RIS, and an interactive page that opens offline. Captions and a methods paragraph are generated in Chinese and English.

Samples

Three maps from open data

The example projects use open data from OpenAlex. Open them in the studio.

Deep learning in medical imaging

OpenAlex open data. The keyword map in the hero is this sample.

Carbon neutrality and emissions trading

OpenAlex open data on policy and carbon-market papers.

Large language model agents

OpenAlex open data on the agent literature.

Open in the studio

How it works

Three steps to a map

From an export file to a figure and a review.

↑

Drop an export file

.txt .csv .ris .net

or search OpenAlex

01Import

Drop a Web of Science, Scopus (CSV), CNKI, Wanfang, CSSCI, or PubMed export, or search OpenAlex.

Hotspots
Intellectual base
Evolution
Collaboration
Run analysis

02Pick a goal

Research hotspots, intellectual base, research front (bibliographic coupling), evolution, collaboration, or core journals and authors. Run one, or open the parameters.

SVGPDFPNGCSVGraphMLGEXFPajekRISAI review

03Read, export, write

Cluster map, timeline, timezone, and bursts sit together. Export vector figures and tables. AI writes a Markdown review with citations.

Why web

SciGalaxy vs the CiteSpace desktop app

Algorithms follow CiteSpace (Chen, 2006). Simpler than CiteSpace.

FileDataNetwork

Java

CiteSpace desktop

  • × SetupInstall Java
  • × Chinese dataConvert it first
  • × LanguageEnglish
  • × AIBring your own OpenAI key
  • × AnalysisTune parameters by hand
scigalaxy.com

SciGalaxy

  • ✓ SetupOpen the browser
  • ✓ Chinese dataImport it directly
  • ✓ LanguageChinese / English
  • ✓ AIBuilt in
  • ✓ AnalysisOne-click runs

Pricing

Pricing will be announced. Local analysis is free to try. AI features (naming, summaries, and the review) and OpenAlex search are billed in credits.

Local analysis

Free to try. Import, networks, clusters, and time views run in the browser.

AI features

Naming, summaries, and the review use credits.

OpenAlex

Online search uses credits.

Start free

Questions

Is my data uploaded?

The analysis runs locally in your browser. The original file is not uploaded. AI features send only the analysis summary.

Will the numbers match CiteSpace?

The algorithms follow CiteSpace. Results agree overall. A few indicators may use a different definition.

CNKI exports have no cited references. What then?

Without references you cannot build document co-citation. Keyword, author, institution, and collaboration networks still run on the CNKI file. For co-citation, use Web of Science, Scopus, or OpenAlex.

Can I use the figures in a paper?

Yes. Export SVG, PDF, PNG at 300 or 600 dpi, and the tables. Captions and a methods paragraph are generated in Chinese and English.

Can the AI invent papers?

Cluster summaries cite only works present in your data. The review has 7 chapters and downloads as Markdown.

Can I move a CiteSpace project over?

Yes. Import the CiteSpace project folder: the data, the parameters, and the alias and exclusion lists.

Make the map for your review

Start from an export file.

Start a free analysis

© 2026 SciGalaxy · scigalaxy.com

PrivacyTermsTutorialContact