Visualizing Patterns and Trends in Scientific Literature – What’s next? by Chaomei Chen

Originally published February 1, 2005

Many of us are interested in visualizing patterns and trends in scientific literature. It can be very exciting and revealing as well as challenging and frustrating. More often than not, a visualized ‘big picture’ of a scientific field invites more questions and more specific needs. Some may want to see more details; others may prefer a birds-eye view.

There are quite a few unanswered questions. I’d like to line up a couple of them here. First of all, given any visualization of scientific literature, who would be able to understand what it is about? If there is such a thing as a typical viewer, what would be the viewer’s knowledge structure? The intended audience of the graphical message carried by the Pioneer spacecraft was aliens who would have competent knowledge of physics, at least as the way we understand it. If designers do not spell out their intent, where are the clues?

The second question may help us to narrow down the answers to the first one. How would seeing an algorithmically visualized world change us? We could become somewhat wiser, somewhat more knowledgeable, or even somewhat more confused. Some changes could be profound and intriguing, whereas some could be superficial and transient. Given the holistic view of science as a whole, how do we measure the short-term as well as long-term impact size of such a revelation?

The third question is about the value of a visualization artifact. Is a naturalistic visualization more valuable than a filtered and synthesized one? Is a prescriptive visualization more desirable than a descriptive one? Is there a non-visual alternative that could bring us straight to the point? In the long run, do we expect to change the way we are thinking, with or without abstract roadmaps of scientific literature?

I’d like to invite you to experience a particular type of visualization – knowledge domain visualization – in CiteSpace. CiteSpace is a Java application that takes bibliographic data retrieved from the Web of Science and visualizes the salient structural and temporal patterns in networks of co-cited articles. The goal is to help us to find out landmarks in a field and how these landmarks are connected. The assumption is that these patterns can help us to get a grip on the dynamics of a scientific field at a macroscopic level. CiteSpace is freely available to anyone, along with a quick user guide. Visit:

The image below is generated by CiteSpace, showing a filtered network of co-cited articles in social network analysis. Can you guess what it is telling?

Figure 1 – A filtered and enhanced network of 721 co-cited articles in social network analysis. The colors are time coded from 1993 (blue) to 2004 (red).

Chaomei Chen, Editor in Chief, Information Visualization
College of Information Science and Technology, Drexel University