Day 50: Kicka$$ science map for your viewing pleasure

Credit: Johan Bollen et al., PLoS ONE

These are the themes of scientific research in more than 2,000 scientific publications, mostly scholarly journals. Pink and blue indicate physics and chemistry, green stands for biology, red represents medicine, and yellow and white indicate social sciences and humanities. 

Johan Bollen, of the Los Alamos National Laboratory in New Mexico, was lead author on a paper that translated Internet user data into this dizzyingly pretty science map. The paper appears in the open access journal PLoS ONE, and I’ll explain a bit about the study below. But first: What are you looking at?

Think of a wheel, where the hub comprises tightly connected social sciences and humanities journals (white, yellow and gray). Secondary classifications for the journals in this cluster include, for example, international studies, religion, music, anthropology, philosophy and finance. The wheel’s outer rim results from a myriad of connections between journals in the natural sciences (red, green, blue): chemistry, hydrology, biology, nursing, pharmacology, etc.

The connections between the journals in the map’s rim cross multiple domains. For example, alternative energy (2:30 p.m., if this were a clock) connects to pharmaceutical research and chemical engineering, which itself further connects to toxicology studies and biotechnology. Brain research (5:30 p.m.) is connected to genetics, biology, animal behavior, and social and personality psychology. Human geography studies connects to geography, plant genetics, and agriculture. And so on.

The connections were all based on Web users’ actions of clicking between studies in various online journals.

Past studies have analyzed the dissemination of science — and the relationships between types of science — based on when and where scholarly articles are cited by subsequent articles. But that method tends to overrepresent the natural sciences, as opposed to the humanities, the authors say. Furthermore, they say, the citation method suffers from lag times between the submission and publication of articles in many journals. And they say it fails to grasp the increasing amount of scientific information that’s being consumed online — and so, it ignores people who are reading scientific studies but not actually citing them in studies of their own.

Bollen and his team collected nearly a billion user interactions at scholarly Internet portals (for example, my accessing this study at PLoS ONE would probably count) during the years 2006 and 2007.  They reason that Web data, which are readily available, “reflect the activities of a broader scholarly community” and allow for the study of scholarship “in real-time, not with a multi-year delay, as is currently the case with citation data.”

Maps of science constructed from online access patterns , the authors say, “can offer an immediate perspective on what is taking place in science and can thus aid the detection of emerging trends, inform funding agencies, and aid researchers in exploring the interdisciplinary relationships between various scientific disciplines.”

I have my questions about some of the details. For example, wouldn’t you expect that readers of minerology studies would also migrate to geology papers? But those two disciplines are at opposite poles, with no connections (rim, 12:30 and 6:30 p.m.). Same with clinical pharmacology and pharmaceutical research (rim, 3 p.m. and 10 p.m.). Perhaps these relationships would become more obvious in repeated trials …

The authors identify their own trouble spots with the new method. They say user privacy is an issue, both from the standpoint of ethics and proprietary interests in the data. And there’s room for bias in the data, they note, as many Web hosts are skilled and ambitious in the generation of “hits,” e.g., visitors, to their sites. 

Still, it’s a good start. And best of all, the new maps look REALLY cool.

You can read the paper here.