This month’s WIRED UK mag­a­zine fea­tures a remix of one of the well-formed.eigenfactor visu­al­iza­tions in their info­porn section.

Together with my col­leagues in Seat­tle and Umea, I mod­i­fied the “change over time” visu­al­iza­tion to tell a spe­cific story: The for­ma­tion of neu­ro­science as a field of its own right over the last decade. Orig­i­nally scat­tered across related dis­ci­plines (such as med­i­cine, mol­e­c­u­lar and cell biol­ogy or neu­rol­ogy), the neu­ro­sci­en­tific jour­nals start to define a niche of their own, reflected in the dense clus­ter emerg­ing in 2005.

Down­load a larger ver­sion with full explana­tory text here: png (1MB) pdf (4MB)

And here is some more in depth info: First, almost 8000 sci­en­tific jour­nals are clus­tered into groups, based on their cita­tion pat­terns, and using the map equa­tion (demo, paper). In short, for a net­work par­ti­tioned into groups, the map equa­tion spec­i­fies the the­o­ret­i­cal limit of how con­cisely we can describe a tra­jec­tory of a ran­dom walker on the net­work. There­fore, min­i­miz­ing the map equa­tion over all pos­si­ble net­work par­ti­tions reveals reg­u­lar­i­ties of infor­ma­tion flow across directed and weighted net­works or, in our case, the struc­ture of how cita­tions flow through science.

Sec­ond, using the Eigen­fac­tor™ Score, the jour­nals are assigned a mea­sure of impor­tance – much as Google’s PageR­ank algo­rithm ranks the impor­tance of web pages. The Eigen­fac­tor™ Score mea­sures the per­cent­age of time that researchers would spend with the respec­tive jour­nal, if they were to move through the net­work by ran­domly fol­low­ing cita­tions in the journals.

This process is repeated in two-year chunks from 1999–2007, in order to cap­ture changes in clus­ter­ing and shifts in impor­tance over the years. For this dia­gram, we picked only the clus­ters rel­e­vant to the for­ma­tion of neuroscience.

In the visu­al­iza­tion, each clus­ter occu­pies a ver­ti­cal col­umn block in the respec­tive year’s col­umn, fur­ther sub­di­vided into a block for each jour­nal. Each jour­nal is con­nected with a hor­i­zon­tal band over the years. The height of each jour­nal reflects the Eigen­fac­tor Score. All jour­nals in the clus­ter that cor­re­sponds to the field of neu­ro­science in year 2007 are high­lighted to tell the story of the for­ma­tion of this field of sci­ence. The col­or­ing is based on the clus­ter assign­ments in the first year, 1999.

We use a sub­set of the cita­tion data from Thom­son Reuters’ Jour­nal Cita­tion Reports 1999–2007. The com­plete data aggre­gate, at the jour­nal level, approx­i­mately 35,000,000 cita­tions from almost 8000 jour­nals over the past decade, but here we only dis­play jour­nals rel­e­vant to the for­ma­tion of neuroscience.