PubTrends is an exploratory tool for researchers providing faster trends analysis and breakthrough papers discovery among the steadily growing flow of papers worldwide. The service aims to solve three tasks: give a brief overview of the field, explore popular trends in publications, and help to find new promising directions.

Open Access Paper:
Poster is available here.
Citation: Shpynov, O. and Nikolai, K., 2021, August. PubTrends: a scientific literature explorer. In Proceedings of the 12th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics (pp. 1-1).



PubTrends contains 30 mln papers and 175 mln citations of biomedical literature from the PubMed® database with 170 mln papers and 600 mln citations from the Semantic Scholar archive. Semantic Scholar aggregates significant journals and publishers, including Springer Nature, ACM, etc.


All the analysis starts with

Next step is known as the Bibliometrics analysis, i.e. citations graph analysis, collecting co-citations (when a pair of papers are cited together), bibliographic coupling (number of common references in a pair of papers), and possible citations based on texts similarity. These values are combined into aggregated similarity score and are used to build similarity graph. Both citations and similarity graphs are available for additional exploration in interactive graph viewer.

Finally, user gets full report covering all the aspects of analysis.


Here we describe the analysis for the search query human aging.
We focus on 1000 most cited papers from the PubMed, with review papers, extending search set with connected papers by 20%.
Learn more exploring one of precomputed search queries from the PubTrends main page.


The report page contains all the analytics and consists of several parts: Papers, Trends, Topics, etc.
Side bar on the left of the page can be used for navigation.
Please use About button - PubTrends will guide you through the report.


The Papers section demonstrates a birds-eye view of the field, including the total number of articles, and extracted topics. Word cloud shows the most frequent words in titles and abstracts. Also, it contains a summary plot of papers per year. Please note that the word cloud component is clickable, and you can navigate to documents containing the selected word. Articles can be viewed as a plain list, as well.
Topics were computed by hierarchical clustering of papers similarity graph, where similarity is computed based on citations information and paper titles and abstracts.

The Trends section contains an interactive visualisation of top-cited papers, organised by number and citations count. Different types of articles are shown in different colours.
Most cited papers and papers with quickest growth of citations are also shown here. All the papers are clickable, and we can explore details on a separate page.


Topics are closely related groups of documents. Similarity between papers is computed based on bibliographic coupling, direct citations, co-citations, and text similarity between papers. Community detection algorithm is used to extract dense clusters of papers or topics. Small topics are merged together to get reasonable number of topics.
Overall structure of topics within a research field can be visualised as a graph.

This graph allows you to find out the most important spots. Hover on any nodes to view paper details, filter papers with dedicated filter field, configure visualization options and more. On the screenshot you can see papers coloured by different topics and the paper is highlighted with its similar neighbours.


In the section Topics for each topic, the application shows familiar to users word cloud and articles plot. Word cloud is built from terms specific to the given topic with respect to others. These words are computed using TF-IDF normalization, a standard approach in the field of natural language processing. The more important word is the more significant fraction of papers contains it.
The topic below is dedicated to DNA methylation changes related with human aging.


Generate a review for the chosen topic - a set of sentences from top cited papers with the highest probability to be included in a real review paper.

Open Access Paper:
Citation: Nikiforovskaya, A., Kapralov, N., Vlasova, A., Shpynov, O. and Shpilman, A., 2020, December. Automatic generation of reviews of scientific papers. In 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA) (pp. 314-319). IEEE.


Other sections include information about the most popular journals, most productive authors, etc.

Paper analysis

In case when user is interested in the particular paper, PubTrends can help user to find similar papers and most important prior and derivative works. User can use title name of DOI for paper identification. The service will search for paper, analyze connected papers and proceed one step further by references to find out which of the connected papers have co-citations, common references, etc. When the set of papers is collected, both citations and similarity graphs are built. User can explore these graphs in integrated viewer.

Your feedback is welcome

Let's make the service better together! We would love to hear your opinion on the results.
Feel free to fill the feedback form on the bottom of reports page or use emotions buttons to share your thoughts.