Friday, May 22, 2009

Each “list of common words unique to each paradigm” forms a streaming ribbon in the image above

TO BE NOTED: From Metamodern:

A Map of Science

by Eric Drexler on May 20, 2009

A comment on my previous post reminded me of a wonderful visualization that amounts to a map of the whole of science, generated by citation-based clustering of almost a million papers. The image above is a view of an extraordinarily information-dense representation, not just of connections among fields, but of their content. At 13,566,672 pixels, most of the text is readable. I have the printed version and examined it with a magnifying glass.

Under the title “A Map of Science”, it was featured by Nature in 2006. Here’s a description by the developers at

As to what the image depicts, it was constructed by sorting roughly 800,000 scientific papers into 776 different scientific paradigms (shown as red and blue circular nodes) based on how often the papers were cited together by authors of other papers. Links (curved lines) were made between the paradigms that shared common members, then treated as rubber bands, holding similar paradigms closer to one another when a physical simulation forced them all apart: thus the layout derives directly from the data. Larger paradigms have more papers. Labels list common words unique to each paradigm.

Each “list of common words unique to each paradigm” forms a streaming ribbon in the image above. What the authors call a paradigm, I would call a field, or topical area (such as seismology, organometallic chemistry, cryptology, virology, and stellar dynamics), clustered within broader areas (such as geophysics, chemistry, computer science, molecular biology, and astrophysics).

At the end of my previous post, I said a bit about what I’ve learned through a huge investment of time in cross-disciplinary study. This is hard to quantify, but it was nice to find that most of the important words specific to each field across the whole of the map are familiar."

No comments: