Overview

What is this tool?

ETO's Map of Science collects and organizes the world’s research literature, revealing trends at the cutting edge of science and technology. The Map organizes hundreds of millions of articles from around the world into over 100,000 clusters, or groups of articles that cite each other often - typically because they share other things in common, such as topic or language. (Learn more about research clusters.) The Map also includes data about each cluster’s growth over time and about the key topics, countries, authors, and funders of the articles within the cluster.

What can I use it for?

You can use the Map to:

• Track trends in research output, growth, and other metrics for science and technology topics from the general to the very specific. The Map’s filters can help you discover and analyze clusters of research about highly precise concepts - everything from robots that grow like vines to speech recognition in noisy environments. Read more >>
• Learn about key countries, researchers, institutions, articles, and concepts in different areas of research using the Map’s detail view mode.
• Identify the research topics with the fastest growth, largest influence, and other measures of impact.
• Build powerful, complex filter combinations with the Map’s easy-to-use interface to quickly zero in on the most relevant topics and actors.

Details and examples >>

What are its most important limitations?

• The Map isn’t built for researching individual people, organizations, or articles. It’s better to look first for particular types of research using the Map’s filters, then browse information about the people, organizations, and articles related to that research - not the other way around. Read more >>
• The Map’s clusters don’t always have a clear meaning. The research clusters in the Map are generated algorithmically using the citation links between articles. In some cases, this method may group articles that don’t have much in common thematically. In most cases, though, studying a cluster will usually show common-sense reasons why the articles in it tend to cite each other (for example, they may share the same language or subject matter). Read more >>
• The Map shows a single point in time. The research clusters displayed in the Map come from ETO’s Research Cluster Dataset, which doesn’t include any information about how individual clusters or fields of research have changed over time, other than some basic growth statistics. Because of this, the Map only gives a “snapshot” view of the research landscape.
• The Map doesn’t cover unpublished research. Its data sources only include research that has been publicly released. We believe these sources cover the large majority of published research worldwide, but we have no way of knowing how much unpublished research exists, or how different it would look if unpublished research were somehow included. More details >>
• Some features are missing or limited in mobile. For the full version of the Map, we recommend using a desktop browser.

For more caveats about the data used in the Map, refer to the documentation for the source datasets.

What are its sources?

The Map of Science relies on ETO’s Merged Academic Corpus and Research Cluster Dataset. Read more >>

Does it contain sensitive information, such as personally identifiable information?

No, other than the names of some prominent researchers (all taken from public documents).

What are the terms of use?

The Map of Science is subject to ETO’s general terms of use. If you use the tool in your own work, please cite us. The underlying datasets are not publicly available due to licensing restrictions.

How do I cite it?

Please cite “ETO Map of Science” and include a link to the tool.

Using the Map

How do I use it?

These instructions focus on the desktop version of the tool. Some features may be missing or act differently on mobile devices.

Viewing the Map

The Map of Science tool has three different views:

• Map view, which is the default for the desktop version. Map view isn’t available on mobile devices.

• List view, which is the default on mobile.

• Cluster detail view, which can be opened by clicking on a specific cluster in Map View or List View.

Map view

If you’re using a desktop browser, opening the Map will bring you here first. This view is built around a visualization of 100,000+ research clusters, each represented by a single dot on the map. Clusters spaced closer together have more inter-cluster citation links (that is, citations by articles in one cluster to articles in the other). The color of each dot indicates the most common general subject category (for example, biology, social sciences, or medicine) among articles in the cluster from the last five years; larger dots represent clusters with more papers added in the past five years.

There’s also a left-hand filter menu, and as you hover over individual clusters in the map, summary details will appear on the right.

To learn more about what the filters mean, hover over the “?” icons.

As you apply filters, clusters that don’t match them will disappear from the map, which will automatically rescale around the ones that remain. If you’d prefer to keep all of the clusters visible, use the “Display Unselected” switch.

Click and drag to zoom in on a portion of the map. To access more controls, navigate to the top right corner of the window.

Clicking on a cluster in the map will bring up its detail view.

As you use the tool, your browser’s address bar will update to reflect your filters and selections. Copy the URL in order to return to the same view later.

List view

The list view shows each cluster as a row in a table. Use the left-hand filter pane to narrow the table to clusters that meet your filters. To learn more about what the filters mean, hover over the “?” icons.

You can add or remove columns with the “Add/Remove Columns” button. Most of the available filters can also be displayed as columns.

By default, the clusters are listed in descending order by growth rate. You can reorder the table by any column with an arrow icon in ascending or descending order.

Click on a cluster to bring up its detail view.

Cluster detail view

Each cluster’s detail view includes information about its growth, concepts, languages, countries, prominent articles, and many other details, all organized into several sections.

Each cluster’s detail view also has a unique URL. Copy the URL in order to return to the same view later.

What can I use it for?

Explore research on any topic, from the general to the very specific. The Map includes searchable filters for hundreds of high-level concepts and fields of study. Tinkering with these filters can lead you to clusters of research about highly precise concepts - everything from robots that grow like vines to speech recognition in noisy environments.

Learn about the key countries, researchers, institutions, and articles in specific areas of research using the Map’s browse and detail view modes.

Identify the research topics with the fastest growth, most citations, and other measures of impact.

Build powerful, complex filters to quickly zero in on the most relevant topics and actors with the Map’s easy-to-use filtering interface.

How have others used it?

Since the Map’s launch in October 2021 as a CSET prototype, researchers at CSET and elsewhere have used it to:

• Identify prominent countries, researchers, and subtopics within specific fields of research, such as superconductor electronics, natural language processing, AI safety, supercomputing, and bioinformatics.
• Explore how AI research priorities differ between top researchers in the U.S. and China.
• Identify the leading countries in overall research output and in specific topics, such as medicine, physics, and computer science.
• Build a general method for tracking progress across different subfields in AI.

Beyond these public examples, users in and out of government consult the Map in decisions related to science and technology policy, diplomacy, industrial policy, national security, and other domains.

We’ll add new public examples here as we learn about them.

Which uses are not recommended?

Researching individual people, organizations, or articles. You can’t search the Map for individual people, organizations, or articles. The Map does include some of this information in the detail view for each cluster, to provide context to that cluster, but in order to view it, you have to have first selected that cluster for other reasons (for example, because it meets criteria you specify with filters). In other words, it’s better to first look for particular types of research using the Map’s filters, then browse information about the people, organizations, and articles related to that research - not the other way around.

Tracking trends over time in the composition of a specific field or cluster of research. The Map provides a current “snapshot in time” of research. You can see which fields have grown fastest overall (as of the time of that snapshot) using the Map’s growth filters, but you can’t currently see how an area of research has evolved over time in terms of participants or funders (for example). It may be possible to approximate some such analyses with careful combinations of filters and cluster details, but the Map isn’t currently designed to support these analyses.

Studying topics unlikely to show up in public literature, such as classified government research or commercially sensitive R&D. The Map relies entirely on public documents, such as published articles, preprints, and conference papers and may not give a good view of other, non-public types of research.

Sources and methodology

All of the data used to generate the Map of Science comes directly from the Research Cluster Dataset, which is itself based on the Merged Academic Corpus. These datasets aren’t publicly available, but you can learn more about them on their documentation pages:

• The Research Cluster Dataset uses data from our Merged Academic Corpus to create over 100,000 citation-based clusters of articles and generates extensive metadata about each one, including cluster growth, language, subject matter, funding, and other topics. All of the visualization, filtering, and other features in the Map of Science use the cluster-level metadata in the Research Cluster Dataset. To understand the "backstory" behind a particular Map of Science feature or filter, refer to the relevant part of the Research Cluster Dataset documentation.
• The Merged Academic Corpus (MAC) contains detailed information on over 270 million scholarly articles, combining data from public and private sources to achieve an unmatched view of the global research literature. The article-level metadata in the MAC is used to build and characterize clusters in the Research Cluster Dataset, meaning that the Map of Science (which relies on the Research Cluster Dataset) is indirectly based on the MAC.

Maintenance

How is it updated?

The Map interface is updated from time to time as new features are developed. The underlying cluster structure and metadata are updated roughly quarterly; see the Research Cluster Dataset documentation for more details.

How can I report an issue?

Use our general issue reporting form, or click on the “Submit feedback” icons embedded in the tool to report issues related to specific data points.

Credits

Many of the algorithms and metrics used in the underlying corpus were developed by SciTech Strategies. Dewey Murdick conceived of and started the Map of Science effort and has informed the design and initial requirements. Ilya Rahkovsky designed and implemented an initial Tableau interface that informed the design of the first version of the Map of Science, as well as taking a leading role in clustering the underlying dataset.

Jennifer Melot implemented the Map of Science user interface and contributed to its current design in collaboration with Zach Arnold. Dewey Murdick, Catherine Aiken, Autumn Toney, and Sara Abdulla have contributed ongoing feedback and ideas on the user interface. Neha Singh, Brian Love, Ashwin Acharya, Catherine Aiken, Zach Arnold, Miguel Camargo, Jack Corrigan, Brian Hayt, Patrick Lee, Maya Mei, Igor Mikolic-Torreira, Mina Narayanan, Avital Percher, George Sienawski, Alex Stevens, Dan Stubbins, Helen Toner, Bryan Ware, and Molly Wasser provided review and testing.

Zach Arnold wrote this documentation.

Major change log

10/6/21	Initial release (CSET prototype)
10/19/22	ETO release